# Project Title Project Abstract

Blockchain technology has emerged as a promising solution for secure and decentralized
data storage and replication. Smart contracts, self-executing programs that run on
blockchain networks, offer a novel approach to automate and manage data replication across
multiple nodes in a blockchain network. This abstract proposes a concept whereby smart
contracts are utilized to facilitate the replication of data across distributed nodes, ensuring
redundancy, fault tolerance, and data integrity.

The proposed concept leverages the unique features of blockchain, including its distributed
ledger, cryptographic hashing, and consensus mechanisms, to establish a trustless and
tamper-resistant environment for data replication. Smart contracts are deployed to automate
the replication process, defining rules and conditions for data transfer and storage. These
Leveraging Smart contracts execute predefined instructions based on specified triggers or events, such as the
Contracts for Data addition or modification of data.
1 Replication Across
Multiple Nodes in
By utilizing smart contracts for data replication, organizations can achieve several benefits,
Blockchain Networks
including enhanced security, transparency, and auditability. The immutable nature of
blockchain ensures that replicated data remains tamper-proof and auditable, providing
assurance of data integrity. Furthermore, the decentralized nature of blockchain networks
offers resilience against single points of failure, enhancing the reliability and availability of
replicated data.
In conclusion, the concept of using smart contracts for data replication in blockchain
networks presents a novel approach to address the challenges of traditional centralized
replication methods. By harnessing the power of blockchain and smart contracts,
organizations can establish a robust and efficient framework for securely replicating data
across distributed nodes, paving the way for decentralized and trustless data management
systems.
School of Computer Science in-house Projects
(Available for 2022, 2023 Intake Students)

Date: 3rd June 2024

School of Computer Science and Engineering

1 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In this project, we aim to develop a secure Internet of Medical Things (IoMT) infrastructure,
which is the foundation of the smart healthcare environment, that will benefit from the latest
advancements in the field of Internet of Things (IoT) and IoMT. Our idea of smart
healthcare is multi-faceted, where every patient is provided wearable devices capable of
capturing required information after a defined time interval and sharing it with the nearest
server module. The proposed system is very helpful for doctors and paramedical staff as it
will provide 24/7 monitoring of patients, which is very important if a patient is in critical
condition. Secondly, it will be helpful for the patients as well, as it will provide timely
information about the intake of a medicine. Thirdly, a smart healthcare infrastructure will be
able to adjust the rooms’ temperature according to their needs (unlike existing systems,
where each room in a building has its own separate air conditioning or heating devices,
which consume a large amount of electricity and gas). Additionally, the security system will
Secure Internet of
also be enhanced, where everyone will use an RFID or biometric entry and exit system. This
Medical Things
2 security system will be connected to the CCTV, and both will communicate in real-time
Infrastructures for
with the central command system/server. It will not only help to bar any intruder or any
Smart Healthcare
unwanted person from entering the smart hospital but also keep a record of each person or
vehicle in the coverage area. Further, the fire system will be integrated with the server, and a
panic alert device will be installed in each ward or apartment. This will help to early warn
people in case of fire or any unwanted situation. Also, to further minimize the exposure of
the inside of the smart hospital to temperature during the summer, a special type of glass
commonly known as smart glass will be installed on those sides of the building that are
exposed to direct sunlight. It will not only help to reduce heat gain but also generate
electricity. The generated electricity can be added to the central energy system and smart
energy meters. To further reduce energy consumption, the lighting system will be connected
to the motion sensors, which will turn on only when it senses motion. This will also help
trace out any illegal activity. This complete IoMT-enabled ecosystem of smart healthcare
has the potential to be adapted for a large-scale deployment in the India.

School of Computer Science and Engineering

2 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Abstract:
This research survey delves into recent publications and future enhancements in Smart
Energy Management Systems (SEMS), with a focus on optimizing power distribution and
consumption through the integration of Artificial Intelligence (AI) techniques. Through an
extensive review of cutting-edge literature and emerging trends, this study identifies key
algorithmic solutions poised to revolutionize energy management practices in the near
future.
Recent publications highlight the importance of rigorous test analysis based on various
parameters to evaluate the performance of AI-driven SEMS. Parameters such as accuracy,
scalability, robustness, computational efficiency, and interpretability are crucial
considerations in assessing the suitability of algorithmic solutions for real-world
deployment.
Smart Energy
Among the diverse array of AI algorithms investigated, Reinforcement Learning (RL)
Management
emerges as a promising solution for optimizing power distribution and consumption within
Systems: Optimizing
3 SEMS. RL algorithms, particularly Deep Reinforcement Learning (DRL) models, offer the
Power Distribution
capability to autonomously learn optimal energy management strategies by iteratively
and Consumption
interacting with dynamic environments. By leveraging real-time data from smart meters,
with AI
weather forecasts, and grid conditions, DRL-based SEMS can adaptively adjust energy
distribution and consumption patterns to maximize efficiency while ensuring grid stability.
Furthermore, recent publications emphasize the potential of Machine Learning (ML)
techniques, such as Gradient Boosting Machines (GBMs) and Long Short-Term Memory
(LSTM) networks, in forecasting energy demand and generation. ML models, trained on
historical consumption data, weather patterns, and socio-economic indicators, can generate
accurate predictions to facilitate proactive energy management decisions, thereby
minimizing wastage and optimizing resource allocation.
In addition to predictive analytics, advanced anomaly detection algorithms, including One-
Class SVMs and Autoencoder-based approaches, are highlighted in recent research for
enhancing the reliability and resilience of SEMS. These techniques enable early identification
of irregularities or malfunctions in power distribution networks, thereby mitigating potential
risks and ensuring uninterrupted operation.
Moreover, recent publications underscore the importance of integrating AI with Internet of
Things (IoT) devices and Distributed Energy Resources (DERs) to enable real-time
monitoring and control of energy assets. By harnessing AI-driven control algorithms, SEMS
can optimize the operation of distributed energy resources, such as solar panels, wind
turbines, and energy storage systems, to enhance grid performance and address peak demand
challenges.
Conclusion:
This research survey highlights the transformative potential of AI in optimizing power
distribution and consumption within Smart Energy Management Systems. Through rigorous
test analysis and evaluation based on key parameters, stakeholders can identify the most
suitable algorithmic solutions for their specific needs. By leveraging RL-based optimization
strategies, ML-driven demand forecasting, advanced anomaly detection techniques, and IoT-
enabled monitoring, SEMS can achieve unprecedented levels of efficiency, reliability, and
sustainability in energy management, paving the way for a smarter, greener, and more
resilient energy future.

School of Computer Science and Engineering

3 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Design and
development of a Planning and management of drinking water ay various floors of the institute is required to
4 decision support be dealt efficiently. Prompt messaging and communication is required from the consumer to
system for water producer using various forecasting and prediction techniques. Therefore, water will be
resource management readily available and future demands and requirements can be forecasted beforehand.

Melanoma, the most deadly and lethal kind of skin cancer, begins on the skin's surface and
progresses deep into the skin's layers. It is one of the cancers that spreads most quickly in
humans. Early detection is essential since it raises the survival probability of melanoma
patients to while enabling them to get straight forward and affordable therapies. This
research provides a novel strategy that uses machine learning techniques for the skin
lesion identification and partitioning into benign or the malignant instances to address these
issues and assist dermatologists. The recommended strategy integrates image processing
methods such as unique contrast stretching of dermoscopic images based on mean values
and standard deviation, as well as image segmentation utilizing the OTSU thresholding
algorithm. Further more, the segmented pictures are used to extract pertinent
Enhancing Melanoma
5 characteristics as color identification features, histogram of oriented gradients (HOG)
Detection Using
features for object recognition, and Texture identification using (GLCM) Gray level Co-
Taxonomy And
occurrence Matrix characteristics. To reduce dimensionality, principal component analysis
Severance
(PCA) is used, and the class imbalance issue is handled with the Synthetic Minority
Oversampling Technique (SMOTE). When contrast stretching is used prior to segmentation,
excellent segmentation results are obtained. Additionally, when using the Random Forest
classifier in conjunction with wrapper-based feature selection, promising results are
obtained when compared to other prominent classifiers. This research contributes to the
advancement of melanoma detection, offering potential cost-effective and accurate solutions
to aid early diagnosis and support healthcare professionals in providing timely and informed
treatments for patients. This innovative approach empowers healthcare professionals to
administer timely and informed treatments, marking a pivotal step towards enhancing
patient outcomes.

School of Computer Science and Engineering

4 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 StreamHarvest promises to automate the smooth transfer of songs from Spotify to consumer
devices, thereby revolutionizing the way we listen to music. With the use of sophisticated
StreamHarvest: AI features, the system seeks to provide consumers with a customized and emotionally
Automating Music impactful listening experience. Through the integration of sophisticated algorithms,
6
Transfers from StreamHarvest endeavors to gain a comprehensive understanding of user tastes, offering
Spotify to MP3 customized music recommendations and emotional evaluations. This initiative presents
itself as a comprehensive solution that goes beyond the constraints of traditional music
streaming services and offers users a more personalized and engaging aural experience.

School of Computer Science and Engineering

5 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Food spoilage in distribution system warehouses in India is a significant concern, impacting
both food security and economic sustainability. Several factors contribute to food spoilage
in these warehouses. Many warehouses in India lack proper infrastructure for storage,
including inadequate temperature control, ventilation, and pest control measures. These
conditions can accelerate food spoilage by promoting mold growth, insect infestations, and
Autonomous
bacterial contamination. Substandard or damaged packaging materials fail to protect food
detection of food
items from external contaminants, moisture, and pests, leading to spoilage. Limited or
7 quality in public
absent monitoring and control systems contribute to delayed detection of spoilage issues,
distribution System
allowing them to escalate before corrective actions can be taken. Autonomous detection of
warehouse in India
food quality in public distribution system (PDS) warehouses in India could be achieved
through the Machine learning Approach. The objective of this work is to training machine
learning models to analyze the data collected from sensors and computer vision systems.
These models can learn to recognize patterns indicative of food spoilage or contamination,
enabling autonomous decision-making regarding the quality of food items.

One of the most challenging issues existing in our Public Distribution System (PDS) is
managing uncertainties in demand, customer behavior, and market trends. Such instability
and unpredictability resulted in numerous cases of excess stock when demand declined or a
shortage of commodities when demand for certain goods increased significantly.
Forecasting is critical in most firms for optimizing and enhancing organizational efficiency
Forecasting Rice
and enabling data-driven strategic planning. Predictions of commodity store and demand
Sales Demand in Fair
can be used for efficient inventory control, labor work schedules, and infrastructure
8 Price Shop of India
planning, and are a key technique for most aspects of supply chain enhancement. The main
using Machine
objective of this research is to forecast rice sales demand in a Fair Price Shop (FPS) of India
Learning Approach
by employing a Machine learning approach. This work shows how past demand data can be
used to estimate future demand and how these forecasts impact the PDS. The forecast
models generated from past data can be verified and validated in the real case application
using the Mean Squared Error (MSE) that helps to forecast the demand of inventory
required in FPS.

School of Computer Science and Engineering

6 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Depression is one of the most prevalent mental health disorders in the country and it is on
the rise as one of the most serious health concerns facing by the individuals as well as the
students in various college and universities. Major depressive disorder is one of several
types of depressive disorders. Major depressive disorder (MDD also known as major
depression ) is a mood disorder characterized by the presence of sad, empty, or irritable
mood, with somatic and cognitive changes that negatively affect the individual's capacity to
function[10]. Major depression is responsible for more physical and social dysfunction than
many chronic medical conditions. This sometimes leads to suicidal tendency also. It is
estimated that about 10 to 15 percent of teens are depressed at any given time. Research
indicates that one in every four adolescents will have an episode of major depression during
their college and with the average age of onset being 16-22 years. MDD causes a lengthy
Design and
period of gloom and hopelessness, and may rob the sufferer of the ability to take pleasure in
development of a
activities or relationships that were previously enjoyable. In some cases, depressive
method for predicting
episodes seem to be triggered by obviously painful events such as suicides. The irony is
students’
9 that it is also one of the most treatable disorders, through psychotherapy
transformation for
and/or medication. Yet barely, a one third of the people with depression seeks help or is
Maximum depressive
properly diagnosed.
disorder and suicidal
In this project a dynamic model is developed that will identify the students suffering
tendency
from different levels of MDD and also predict the students transforming serious MDD to
suicidal tendency. In this an online test will be conducted for the students two times per year
with well-defined questionnaire. Then classify the students by who are suffering with
different levels of MDD (i.e. Level1, Level2, and Level3) by using a proposed method fuzzy
decision tree. After that this model will predict by a logistic regression when the students
will transform from Level1 to Level2, Level2 to Level3, and Level3 to suicidal tendency
and purpose appropriate methods to mitigate the MDD risk among the students

Objectives:
1. Developing an appropriate questionnaire for online test which will extract various
levels of MDD risk among the students
2. Pre-processing the data received after online test (i.e. Eliminating missingness and
uncertainties from the data)
3. Classifying the students into different groups based on their levels of MDD (i.e.
Level1, Level2, and Level3)
4. Predicting when the students will transform from Level1 to Level2, Level2 to Level3,
and Level3 to suicidal tendency
5. Suggest appropriate actions to be taken at each level (Post Processing)

The growth of electronic health records (EHRs) holding essential medical images has
Securing Medical
Images in Cognitive prompted a demand for strong data protection and quicker diagnostic procedures in the fast
IoT Networks with changing world of healthcare. In response, this research presents a unique, security-based
10 SCSOA based cognitive IoT network that uses a secure medical data management architecture driven by
SIMON Encryption Hyperledger blockchain (HB) technology. This ground-breaking paradigm integrates a

School of Computer Science and Engineering

7 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 on Hyperledger
Blockchain number of crucial processes, such as encryption, optimal key generation, and safe data
management, which are all coordinated to guarantee the highest levels of data privacy and
access control.

School of Computer Science and Engineering

8 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Artificial Intelligence is the master mind of all the Innovations in the Modern Socio Techno
Life whereas, Machine Learning (ML) is the Mankind of making Smarter Decisions. In our
work by applying Machine Learning with Deep Learning (DL) in an ensemble manner to
diagnose whether a person is affected by Heart Disease or not. By taking in to account,
algorithms like K-Nearest Neighbor (KNN), Random Forest in association with Deep neural
network and Optimized Feature Selection method for the reduction of weights between the
neurons to achieve the Highest level of Prediction. This will lead to categorization of
Disease in a fruitful manner to Understand the diseases in earlier stages for the avoidance
of mortality rates. Here with the help of their prime attributes in an accountability of 14
An Effective Heart counts as features. The proposed Model shows Higher Performance in terms of prediction
11 Disease Prediction metrices such as Classification Accuracy, Precision Rate, Error Rate, Specificity,
Model via Ensembling Sensitivity, F1-Score for Good Decision Making.

The main objective is:

1. To keep a check of the quantity used. So, that the user would get frequent updates about
Automated Grocery the quantity and freshness of the products stored.
12
Ordering System 2. To keep a track of the freshness of the groceries.
3. To develop a software/app which provides information about the quantity of the grocery
depleted and about the freshness of the groceries stored.

School of Computer Science and Engineering

9 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This research investigates the application of ensemble learning techniques for AD
classification using magnetic resonance imaging (MRI) of the brain. The proposed model
integrates the strengths of U-Net, a convolutional neural network (CNN) architecture
renowned for spatial feature extraction in medical images, with AdaBoost, an ensemble
learning algorithm, to enhance classification performance.
Hybrid Learning The study utilizes a curated dataset of brain MRI scans, comprising images from subjects
mechanism for with confirmed AD diagnoses and healthy controls. A U-Net architecture is employed to
Alzheimer's Disease extract intricate spatial features from the MRI images, capturing nuanced patterns associated
detection: Integrating with AD-related structural changes. These features are then fed into an ensemble of weak
13
U-Net Feature classifiers, specifically decision trees, orchestrated by the AdaBoost algorithm. AdaBoost
Extraction with adapts by assigning higher weights to misclassified instances, iteratively improving the
AdaBoost. ensemble's performance.

A realtime project in which students will develop the following modules - Leave application
14 JAIN UMS - HRM and approval module, attendance tracking and report generation, grievance management and
employee announcements, engagement and recognition

In today's world plant diseases are growing rapidly due to the global warming. The effect of
climatic changes have the significant impact on the health of both human's and plants. In
Drumstick leaf this project we mainly concentrate on the specific types of diseases that can spread over the
15 disease detection plants. Particularly the examination of disease in the leaf part of the drumstick plant is
using GAN throughly detected by using diverse of ML algorithms and the detection is improvised usind
GAN. Diverse ML models are evaluated to discover their performance with the highly
sophisticated GAN model.

School of Computer Science and Engineering

10 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Steganography is the practice of concealing messages or information within other non-secret
text or data. This research work explores an innovative approach to enhance data hiding
capacity within the framework of code-based steganography. Traditional steganographic
methods often face challenges related to limited data capacity and vulnerability to detection.
To address these issues, we propose a novel technique that integrates error-correcting codes
with advanced encoding algorithms, significantly increasing the amount of data that can be
embedded while maintaining the integrity and security of the hidden information. The
proposed method utilizes a combination of low-density parity-check (LDPC) codes and
Huffman coding to optimize the encoding process. LDPC codes are employed for their
An Efficient and robust error-correcting capabilities, ensuring that the hidden data remains intact even in the
Enhancing data hiding presence of noise or compression. Huffman coding is used to efficiently compress the secret
16
capacity in code- data before embedding, maximizing the use of available space within the cover medium.
based steganography. Extensive experiments and analyses demonstrate that our technique offers a substantial
improvement in data hiding capacity compared to conventional methods. Furthermore, the
enhanced security measures reduce the likelihood of detection by steganalysis tools. The
results show that the proposed approach not only increases the embedding capacity but also
maintains a high level of fidelity in the cover medium, making it a viable solution for secure
and efficient steganographic applications. In conclusion, this work presents a significant
advancement in the field of code-based steganography, offering a method that balances
increased data hiding capacity with enhanced security features. The integration of LDPC
codes and Huffman coding sets a new benchmark for steganographic techniques, paving the
way for future research and practical implementations in secure communications.

Bone fractures are a common injury that requires accurate and timely diagnosis for effective
treatment. This project aims to develop a deep learning-based system for detecting bone
Deep Learning-Based fractures in X-ray images using a convolutional neural network (CNN). Utilizing the Bone
Bone Fracture Fracture Multi-Region X-ray dataset from Kaggle, the system will be trained to identify and
17
Detection on X-ray classify fractures across different regions of the body. The proposed model will enhance
Images diagnostic accuracy and speed, assisting radiologists in identifying fractures more
efficiently and reliably. By automating the detection process, the system aims to reduce
diagnostic errors and improve patient outcomes.

School of Computer Science and Engineering

11 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Cloud computing enables scalable and on-demand access to virtualized resources, but
efficient resource allocation remains challenging due to dynamic workloads, heterogeneous
resources, and varying quality of service (QoS) requirements. Traditional allocation
techniques often struggle with the complexity of cloud environments. This project proposes
to leverage reinforcement learning (RL) for intelligent and adaptive resource allocation in
Resource Allocation clouds. RL techniques allow learning optimal policies through trial-and-error interactions
in Cloud using with the environment, making them well-suited for the dynamic cloud context. The research
18
Reinforcement methodology involves formulating the resource allocation problem as a RL problem,
Learning designing RL algorithms like Q-learning or deep RL tailored for load balancing, energy
optimization and QoS management, implementing the algorithms, and evaluating
performance through simulations and comparative analysis against traditional approaches.
Expected outcomes include novel RL-based allocation algorithms, performance
benchmarking, insights into real-world applications, and academic publications advancing
RL for efficient cloud resource management.

School of Computer Science and Engineering

12 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Real-Time Virtual using hand gestures as an input method for controlling a cursor, eliminating the need for a
19 Mouse using Hand physical mouse. It involves sensors and machine learning algorithms to translate gestures
Gestures into cursor movements.

School of Computer Science and Engineering

13 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Traffic accidents are a significantly growing global concern nowadays, causing fatalities
and injuries. Therefore, to improve road safety and develop a smart traffic management
system, a framework for vehicle crash prediction is essential for urban and cities.
Traditional traffic safety research often relied on statistical methods for long-term vehicle
crash prediction using historical data e.g., yearly or monthly traffic data. While these
Integrative methods do not endorse complex relationships between multiple factors that contribute to
Framework for Smart crash prediction. Recently, there has been a shift in the research spectrum towards short-
20 Road Safety: Traffic term crash prediction frameworks using Artificial Neural Networks (ANNs). These include
Accident Modeling various ANN architectures like single-layer perceptrons, multi-layer perceptrons,
and Prediction convolutional neural networks, and recurrent neural networks. These models have produced
an acceptable level of accuracy in forecasting short-term traffic crashes but lagging in terms
of model complexity and execution time. To further enhance prediction performance, this
work proposes a framework for traffic modeling and prediction that exploits Convolutional
Recurrent Neural Network (CRNN). This approach employs both temporal (time-based) and
spatial (location-based) features of traffic data for improved crash prediction.

This project focuses on developing a smart home energy management system using machine
learning algorithms and IoT devices. By analyzing energy consumption patterns, user
Smart Home Energy preferences, and external factors such as weather conditions, the system will optimize energy
21 Management System usage, reduce utility costs, and promote energy efficiency in residential settings.
using ML

Problem: Prevention is always better than cure. And when it comes to fire, preventive
measure is an absolute necessity Fire is a destructive force that can quickly spread and
inflict unthinkable harm to property, health, and even lives. That is why we should install
fire safety components in the household and the work environment.

Smoke Detecting IoT Solution: Smoke Detecting IoT device is a smart fire detection system that can detect
Device Using Gas combustible gases and alert you to act immediately to control or stop the fire from breaking
22
Sensor – IoT Home out. With the help of Arduino, an MQ-2 Smoke detection sensor, a breadboard, some
Safety jumper wires, a resistor, two LEDs, and a buzzer, one can quickly build this fire detection
system using IoT.

This IoT Project Can Be Very Helpful For

Home or work environment safety

School of Computer Science and Engineering

14 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 School of Computer Science and Engineering
15 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The development of sophisticated conversational AI systems—like chatbots and virtual
NLP for Kannada
assistants—that can comprehend and react to natural language inputs will be greatly aided
23 language chatbot
by NLP. In order for this to function as a virtual assistant for the Kannada language, more
creation using AI
data sources must be developed, trained, and then used.

:Many unknown health issues are being currently faced by different kinds of people. To
overcome from the death rate identification of diseases at current time is essential. All
medical data can efficiently compute and compare in supporting platform of Virtualization
in Big data.The stored data can be mined by fuzzy rule with expert system for
prediction of future health condition. The analysis and prediction of data can be carry out by
using neuro fuzzy algorithm with genetic algorithm. The neuro fuzzy algorithm which
optimize the weights needed for efficiently calculating trained input values in inference
Predetermined system.The EHR(Electronic Health Records) of patient status provided
24 malicious application to the doctor and user by means of automated with approval of expert system. The proper
using BDSA approach security mechanisms is needed for data protection.The hackers trace and may modify the
medical data also. Here the extraction of attack
features performed through graph based event correlation and map reduce parser based
identification method.Attack presence is determined by machine learning like belief
propogation and logistic regression.This two method can detect how many times hackers
tried to hack the data and the details of system trough the hackers
hacked all are saved in server. From the details prevention of future based hacking data can
also be possible

Presenting a comprehensive approach to predicting health status using machine learning

(ML) techniques on streaming data. The proposed system integrates various ML algorithms
to process and analyze continuous health data streams, enabling early detection of potential
Health Status
health issues. By leveraging real-time data processing frameworks and advanced ML
25 Prediction on
models, our approach aims to provide timely and accurate health status predictions,
Streaming data
facilitating proactive healthcare interventions. The system's efficacy is demonstrated
through extensive experiments and evaluations on real-world datasets, highlighting its
potential to enhance personalized healthcare and improve patient outcomes.

School of Computer Science and Engineering

16 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In today's interconnected digital landscape, cybersecurity threats continue to evolve in
complexity and sophistication, posing significant challenges to traditional defense
mechanisms. This research project aims to explore the intersection of Artificial Intelligence
Intelligent Network (AI) and Machine Learning (ML) techniques within the cybersecurity domain, specifically
Defense: Leveraging focusing on network security. By harnessing the power of AI and ML algorithms, this study
Machine Learning for seeks to develop innovative approaches to detect, prevent, and respond to cyber threats in
26
Enhanced real-time. Through comprehensive data analysis, anomaly detection, and predictive
Cybersecurity in modeling, the research endeavors to enhance the resilience of network infrastructures
Complex Networks against a wide range of cyber attacks, including malware, intrusions, and insider threats.
The project will not only contribute to advancing the state-of-the-art in cybersecurity but
also provide practical insights for designing more robust defense strategies in an
increasingly interconnected world.

In order to analyse a variety of cat picture datasets, this study investigates deep learning
approaches. Convolutional neural network (CNN) architecture is trained after preprocessing
Comparative analysis
procedures verify data compatibility. Classification accuracy is improved using transfer
of the activation
learning from pre-trained models such as VGG16 or ResNet. Accuracy, precision, recall,
27 functions with CNN
and F1-score are among the evaluation measures; confusion matrices show how well the
for Cats Image
model performs across different cat breeds. The results show how well deep learning works
Dataset Analysis
for identifying cat photos, which has ramifications for a number of fields including
veterinary care and animal welfare.

Improve the Web

Personalization Using Web personalization is kind of method that is applied to modify a web site to suit the exact
28 the Boosted Random needs of the users, achieving the advantage of data accomplished for the understanding the
Forest for Web directional conduct of users concerning inclusion of more materials in the web framework.
Information Extraction

Designing Machine Identify patterns in human behavior using Computational psychology. Computational
learning model psychology is a field that merges principles of psychology with computational techniques
29 assistant for (like machine learning techniques, artificial neural networks etc.) to model and understand
Diagnosing mental cognitive processes and behavior. It involves the use of mathematical models, simulations,
health conditions and computational methods to study the mind and behavior.

School of Computer Science and Engineering

17 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The water footprint measures the amount of water used to produce each of the goods and
services we use. The water footprint helps us understand for what purposes our limited
freshwater resources are being consumed and polluted. The impact of it depends on where
the water is taken from and when if it comes from a place where water is already scarce, the
Use of Digital consequences can be significant and require action. The increase in the amount of non-
Technology to available water due to pollution and scarce groundwater level has added more water
calculate water footprints, at the community as well as at the personal levels. An increased water footprint
30 footprints for directly affects the health and future of the citizens. Preventing severe drought in water-
different daily use stressed areas is only going to be possible if water is used with more care and efficiency,
items.(Clean & Green this can be done if we have readily available data on water footprints. Hence by using digital
Technology) technologies like AI, Big Data, Blockchain, etc, and computer languages, a user-friendly
app or website may be developed which can provide the water footprints of different items/
final products we use in daily life by feeding little inputs or just by scanning through the
camera like Google lens. The app should support local languages, this will ensure pan-India
usage and sensitize the people about the water footprints of items they use in daily life.

Credit card fraud detection using machine learning algorithms involves the systematic
identification of unauthorized transactions to prevent financial losses. This process includes
collecting and preprocessing transaction data, feature engineering, and labeling transactions
Credit card fraud as fraudulent or non-fraudulent. Supervised learning algorithms, such as logistic regression
detection using and neural networks, are employed alongside unsupervised techniques like anomaly
31
machine learning detection to distinguish between legitimate and fraudulent transactions. The models are
algorithms trained, validated, and deployed to monitor real-time transactions, flagging suspicious
activities for further investigation. Continuous learning from new data enhances the model's
accuracy and reduces false positives, thereby improving the effectiveness of fraud detection
and ensuring better security for financial transactions.

This topic presents Vision Assist, an innovative Android application designed to empower
visually impaired individuals by providing real-time text and product recognition
capabilities using a smartphone's camera. Vision Assist aims to bridge the accessibility gap
by leveraging cutting-edge technologies to enhance independence and facilitate efficient
Vision -Android App daily tasks for users with visual impairments. The application integrates Optical Character
Camera For Text And Recognition (OCR) and Object Recognition technologies, enabling users to identify and
32
Product Labelling For comprehend printed text and recognize various products or objects in their surroundings. By
Unsighted People harnessing OCR algorithms, Vision Assist swiftly converts captured text from images into
audible speech, assisting users in accessing information from books, signs, documents, and
more. Moreover, the object recognition feature employs machine learning models to identify
common items, enabling users to independently navigate their environment and identify
products while shopping or exploring unfamiliar surroundings.

School of Computer Science and Engineering

18 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Transaction fraud imposes serious threats on e-commerce shopping. As online transaction is
becoming more popular the types of online transaction frauds associated with this are also
rising which affects the financial industry. This fraud detection system has the capability to
restrict and block the transaction performed by the attacker from genuine user’s credit card
details. To overcome these problems, this system here is developed for the transactions
higher than the customer’s current transaction limit. During registration we take required
information which is efficient to detect fraudulent user activity. The details of items
Online Transaction
purchased by any Individual transaction are usually not known to any Fraud Detection
Fraud Detection using
System (FDS) running at the bank that issues credit cards to the cardholders. BLA
33 Python &
(Behavior and Location Analysis) is implemented for addressing this problem. An FDS runs
Backlogging on E-
at a credit card issuing bank. Each incoming transaction is submitted to the FDS for
Commerce
verification. FDS receives the card details and transaction value to verify, whether the
transaction is genuine or not. The types of goods that are bought in that transaction are not
known to the FDS. Bank declines the transaction if FDS confirms the transaction to be
fraud. User spending patterns and geographical location is used to verify the identity. If any
unusual pattern is detected, the system requires re-verification. Based upon previous data of
that user, the system recognizes unusual patterns in the payment procedure. After 3 invalid
attempts the system will block the user.

In today's digital age, ensuring the confidentiality and integrity of sensitive information is
paramount. Text and image data, being ubiquitous in electronic communication, demand
Text and Image robust encryption techniques to safeguard against unauthorized access and tampering. This
Encryption paper presents a comprehensive approach to secure text and image encryption and
34
Decryption Using decryption utilizing the Advanced Encryption Standard (AES) algorithm. The Advanced
AES Algorithm Encryption Standard (AES) is a symmetric encryption algorithm widely adopted for its
robustness and efficiency. It operates on fixed-size blocks of data, making it suitable for
both text and image

India, with a rich heritage of floral diversity, is well-known for its medicinal plant wealth,
but their identification is one of the major burning issues in Ayurvedic Pharmaceutics.
Several crude drugs are being sold under the same name in the market leading to confusion
Identification of
and their misidentification. Even the collectors and traders are not completely aware of the
Different Medicinal
exact morphological appearance or differentiating attributes of the many drugs owing to
Plants/Raw materials
seasonal and geographical availability, and similar characteristics. Moreover, the extensive
35 through Image
consumption to meet demand-supply ratio exerts a heavy strain on the existing resources. It
Processing Using
further leads to the practice of adulteration, substitution, and disbelief in the curative
Machine Learning
capability of the system eventually. Thus, software capable of identifying different
Algorithms
medicinal plants/ raw materials through Image Processing Using Different Machine
Learning Algorithms will be of immense use. It will be helpful at every level viz.
wholesaler, distributor, etc. of the supply chain of raw material being utilized in the system.

School of Computer Science and Engineering

19 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In recent years, vehicle theft has become a significant concern for law enforcement agencies
and vehicle owners alike. Traditional methods of tracking stolen vehicles often prove
inefficient, leading to low recovery rates and prolonged recovery times. This project,
"Advanced AI Solutions for Real-Time Stolen Vehicle Tracking and Alerts," aims to
Advanced AI revolutionize the approach to vehicle theft detection and recovery through the application of
Solutions for Real- cutting-edge artificial intelligence technologies. The proposed system leverages a
36
Time Stolen Vehicle combination of machine learning algorithms, real-time data processing, and Internet of
Tracking and Alerts Things (IoT) devices to provide a robust solution for stolen vehicle tracking. By integrating
GPS, on-board diagnostics (OBD), and other sensor data, the system can monitor vehicle
movements continuously. Machine learning models are trained to identify anomalous
behaviour’s indicative of theft, triggering immediate alerts to both the vehicle owner and
relevant authorities.

In the current evolving landscape of marketing, a robust

analytics platform is required that integrates data from various
marketing channels for insightful decision-making. This paper
introduces an Integrated Marketing Data Analytics Platform, a
Advanced Data
comprehensive solution designed to enhance marketing
Extraction and
strategies through advanced analytics. The platform
Analytics Software on
amalgamates data from different marketing sources, enabling
37 AWS:
an integrated view of campaigns, customer behavior, and
Revolutionizing
market trends. Key features include cross-channel analytics,
Insights with
predictive modeling, and real-time reporting. This research
Personalized
aims to empower marketing professionals with a versatile tool
that maximizes the impact and outcome of marketing efforts
by providing actionable insights and promoting data-driven
decision-making in marketing strategy.

School of Computer Science and Engineering

20 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In contemporary times, the establishment of effective communication and community
building among university students belonging to Generation Z has emerged as a crucial
element in facilitating the dissemination of market information and fostering robust
networking. The objective of our work is to Create a chat bot and information portal for
Unified university students to facilitate communication among themselves and with alumni. This
communication project will cover the design, development, and deployment of a comprehensive platform.
38
platform for The main scope of our work is to create Chat bot for real-time communication and FAQs.
university students And Information portal for resource sharing, event announcements, and discussion forums.
We will consider the Chat bot with natural language processing (NLP) capabilities,
Information portal with user-friendly interface and Data privacy and security measures. This
project aims to create a robust communication platform for university students and alumni,
enhancing connectivity and resource sharing.

School of Computer Science and Engineering

21 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This project aims to develop a comprehensive flood prediction and risk management system
for metropolitan areas, leveraging machine learning algorithms and real-time data analysis.
With the increasing frequency and severity of floods caused by heavy rainfall, especially in
urban centers, there is a critical need for proactive measures to mitigate risks and ensure
public safety.
Predictive Analytics The proposed system utilizes machine-learning algorithms to analyse historical rainfall data,
for Mitigation of geographical features, and other relevant factors to predict flood occurrences in
Natural Hazards in metropolitan areas accurately. By harnessing the power of artificial intelligence, the system
39
Metropolitan Cities can forecast potential flood-prone areas with high precision, enabling timely intervention
using Edge and proactive measures.
Computing. Furthermore, the system uses edge computing to capture real-time images, which are
analysed to assess the risk level of flood-prone areas. The analysis is then used to generate
alert messages, which are delivered to the local officials and other relevant authorities.
These alerts provide actionable insights and recommendations, enabling officials to
coordinate rescue and relief efforts effectively.

In recent years, the integration of artificial intelligence (AI) into healthcare has
revolutionized the way medical information is managed and how patient care is delivered.
This project aims to develop an AI-based personal health assistant that can provide users
with personalized health recommendations, track their fitness goals, and offer insights into
their overall well-being. The assistant will leverage machine learning algorithms to analyze
AI-Based Personal data from wearable devices, health records, and user inputs to deliver real-time advice and
40
Health Assistant alerts. Key features include symptom checking, medication reminders, diet suggestions, and
virtual consultations with healthcare professionals. By utilizing natural language processing
(NLP), the assistant can interact with users in a conversational manner, making it accessible
and user-friendly. This project will not only enhance individual health management but also
contribute to preventive healthcare by encouraging healthier lifestyles through personalized
recommendations

There are multiple cloud vendors in the market across the globe. A software developer who
Parameter Ranking
need cloud services for their developmental activities has to choose one of the clouds.
Priority Level
Selection of cloud vendor is a challenging task. The algorithm designed here will provide
Weightage based
41 the developer to choose the best suited cloud service provider. As of now the algorithm can
Cloud selection
be used manually, calculate the average weightage and select the best suited CSP. The
algorithm
implementation of PRPLW algorithm and packaged to software can provide the selected
Implementaion.
CSP as output by taking the parameters for development as input.

School of Computer Science and Engineering

22 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This research introduces an innovative approach integrating automatic attendance
monitoring, facial recognition technologies, and parental notification systems. The proposed
method employs facial recognition algorithms to deliver reliable and efficient attendance
tracking, offering a comprehensive alternative to traditional attendance management
methods. The system achieves exceptional facial detection and identification accuracy by
Advanced Student utilizing Convolutional Neural Network (CNN) techniques and 128-dimensional
Attendance System embeddings for image representation. Implemented with Flask, a lightweight Python web
42 with Deep Learning- application framework, and OpenCV, an open-source image processing package, the system
based Based Face provides a scalable and user-friendly solution. The implementation process involves
Recognition identifying faces in webcam photographs and comparing them to images of registered
students. New users can easily register by providing the necessary details and a photo, while
administrators can efficiently manage existing member data. The student classifier achieved
100% accuracy in the experimental test while in an imaging-friendly position. Additionally,
the system automatically notifies parents of their child's attendance status, enhancing
communication and transparency

In the remote medium, the jammer is able to do just attack the Multi-Hop Wireless network
system (MHWNs). The open-access methodology of MHWNS is the principal purpose
behind the attack of jammer. The principle downside of sticking is to stop the
correspondence by connecting with the validated nodes which have a similar remote
medium or channel. It is fundamental to recognize the jammer since it might give the
required data or message to initiate the instruments of hostile to sticking. Existing strategies
are delicate and offer significance to following and following items through the aggregate
design of the system. Despite the fact that this system gives an answer for the jammer
An Optimized attacks it was not successfully supported to stop the restriction inconvenience on the
Mechanism to grounds that the earnestness of the attack appeared during the season of system
Improve the transmission. To conquer this issue, the proposed framework presents a new calculation
Performance Packet called as Optimized Transmission Mechanism (OTM) which includes the degree of node
43
transmission in handover, picking of definite node to go in the system. Also, this proposed mechanism
Jammer Based on diverts the packets which are available in the jammer location to the following accessible
Multi-Hop Wireless area and make the packet transmission successfully. The motivation behind this exertion is
Network to maintain a strategic distance from the intrusion of the system in the event of sticking by
circumvent the jammer zone and distinguishing another course to beat the current issues.
The different ways are favored with the most astounding association quality to proceed with
the system Quality of Service (QoS) even consequent to sticking. This proposed work
clarifies the OTM calculation which gives the course dependent on the radio -transmitter
location way. This OTM calculation works dependent on the nature of the ideal connection.
The exhibition of the proposed method is evaluated by its packet conveyance value. Results
demonstrate that the exhibition of Extended Optimized Transmission Mechanism (EOTM)

School of Computer Science and Engineering

23 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 is impressively improved than the current calculations.

School of Computer Science and Engineering

24 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Smart Traffic management system (STMS) is a one of the important feature for smart city.
Currently traffic management and alert systems are not fulling needs of STMS. It is more
expensive and highly configurable to provide better service for traffic management. This
Real-time smart
paper proposes a low cost Real-Time smart traffic Management System to provide better
traffic management
service by deploying traffic indicators to update the traffic details instantly. Low cost
system for smart
44 vehicle detecting sensors are embed in the middle of road for every 500 meters or 1000
cities by using
meters. Internet of Things (IoT) are being used to acquire traffic data quickly and send it for
Internet of Things and
processing. The Real time streaming data is sent for Big Data analytics. There are several
big data
analytical scriptures to analyze the traffic density and provide solution through predictive
analytics. A mobile application is developed as user interface to explore the density of
traffic at various places

This project aims to develop an advanced recommendation system leveraging large

language models (LLMs) and natural language processing (NLP) techniques. By integrating
Enhancing User
LLMs' robust text comprehension and generation capabilities, the system will provide
Experience with an
personalized content recommendations across various domains, including e-commerce,
NLP-Driven LLM-
45 streaming services, and social media. The LLM will analyze user interactions and
Based
preferences, extracting meaningful patterns and insights to predict and suggest relevant
Recommendation
items. This approach promises to enhance user engagement and satisfaction by delivering
System
highly accurate and context-aware recommendations, setting a new standard in personalized
digital experiences.

School of Computer Science and Engineering

25 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Federated learning (FL) presents a promising approach for collaborative machine learning in
healthcare by enabling model training across decentralized data sources while maintaining
data privacy. This study addresses critical challenges inherent in applying FL to healthcare
data, focusing on privacy and security, data heterogeneity, and communication efficiency.
We propose a comprehensive FL framework designed to enhance patient data privacy
through advanced techniques such as differential privacy, secure multi-party computation,
and homomorphic encryption. To manage data heterogeneity, we introduce algorithms that
effectively handle non-iid and unbalanced data distributions, incorporating robust
aggregation methods and personalized model updates. This advancement promises improved
machine learning models derived from rich, diverse healthcare datasets, ultimately
contributing to enhanced patient care and personalized treatment outcomes.
Enhancing Privacy
and Efficiency in
46
Federated Learning
for Healthcare Data

Alzheimer's Disease is the most prevalent and pre-dominently found in various parts of USA
Recommendation and across the globe. The study of model that recommends, whether the patient is suffering
Model for Detection from any Memory related issues. If so, what is the stage of memory issue the patient is
47 and Prediction of facing. The Recommendation System is a study of various aspects that is causing the
early Stage patients in various phases of Alzheimer's Disease. These causes are addressed with Bio-
Alzheimer's Disease medical Imaging biomarkers. The Medical Imaging biomarkers include various images of
interest like PET, MRI, sMRI and fMRI images for detection and prediction of the same.

School of Computer Science and Engineering

26 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The early diagnosis and continuous monitoring of neurological and neurodegenerative
diseases, such as Alzheimer's, Parkinson's, and Multiple Sclerosis, remain challenging due
to their complex etiology and the subtlety of early symptoms. Recent advancements in deep
Deep Neural learning, a subset of machine learning, offer promising avenues to address these challenges.
Approaches to Early This study presents a novel deep learning framework, NeuroNet, designed to analyze multi-
48 Identification of modal medical data, including MRI scans, PET scans, and patient-generated verbal and
Neurodegenerative written reports. NeuroNet employs a combination of convolutional neural networks (CNNs)
Conditions for image analysis and recurrent neural networks (RNNs) for sequential data interpretation.
Our system was trained on a curated dataset of 10,000 patients, encompassing various stages
of neurological diseases, and achieved an accuracy of 95.2% in early disease detection,
outperforming traditional methods by 8.7%.

BrainBridge, the world’s first concept for a head transplant system, which integrates
advanced robotics and artificial intelligence to execute complete head and face
Brain Bridge
49 transplantation procedures. This state-of-the-art system offers new hope to patients suffering
Technology
from untreatable conditions such as stage-4 cancer, paralysis, and neurodegenerative
diseases like Alzheimer’s and Parkinson’s.

IoT enabled smart EV

50 Smart EV station
charging station.
51 Not yet decided Not yet decided

ENHANCING The fundamental aim of digital forensics is to discover, investigate and protect an evidence,
CLOUD increasing cybercrime enforces digital forensics team to have more accurate evidence
FORENSICS IN handling. This makes digital evidence as an important factor to link individual with criminal
MAINTAINING activity. In this procedure of forensics investigation, maintaining integrity of the evidence
SECURE plays an important role.Because of the distributed nature of cloud computing, the data of
52
EVIDENCE cloud may be stored and processed across jurisdictional borders. In such case, collecting
STORAGE BY digital evidence from cloud computing requires multiparty cooperation. This work proposed
APPLYING a process provenance, which provides proof of existence and privacy preservation for
BLOCKCHAIN process records by using technologies of blockchain and cryptography group signature. The
METHODOLOGY process provenance enhances the trustworthy of the chain of custody for cloud forensics.

A question answering system for university policies would be a natural language processing
University Policy (NLP) application designed to provide accurate and relevant answers to questions related to
53 question answering the various policies and regulations of a university. The system would aim to help students,
system faculty, and staff easily find information about topics such as academic policies, student
conduct, financial aid, housing rules, and more.

School of Computer Science and Engineering

27 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The rapid expansion of digital networks and the increasing sophistication of cyber threats
necessitate innovative approaches to network security. Blockchain technology, with its
decentralized, transparent, and immutable characteristics, offers promising solutions for
enhancing network security. This project explores the application of blockchain technology
to various aspects of network security, including authentication, data integrity, secure
communications, and intrusion detection.

The primary objective of this study is to design and implement a blockchain-based

framework that addresses prevalent network security challenges. The framework leverages
54 Blockchain-Based
blockchain’s distributed ledger to create a tamper-proof record of network activities,
Solutions for Network
Security Enhancements ensuring data integrity and transparency. Additionally, smart contracts are utilized to
automate security protocols and enforce access control policies, thereby reducing the
reliance on centralized authorities and mitigating single points of failure.

This project also investigates the integration of blockchain with existing network security
mechanisms, such as Public Key Infrastructure (PKI), to enhance authentication processes.
By storing cryptographic keys and digital certificates on the blockchain, the framework
provides a more secure and resilient method of managing digital identities.

The process of image description is one of the growing research problem . The model using
different combination
55 Image captioning using Of the neural network along with long short term memory to get better results, finding the
deep learning best combination is challenging based on accuracy, this project work focus on the
algorithms mentioned challenges and developing a better algorithm solve the problem.

Application of
Creating an Intrusion Detection System (IDS) using machine learning algorithms is an
machine learning
exciting and impactful project. Developing an IDS using machine learning involves a
algorithms to improve
56 combination of data science, cybersecurity, and software engineering skills. The project not
the accuracy of
only provides practical experience with machine learning techniques but also contributes to
intrusion detection
the crucial field of network security.
systems (IDS)

School of Computer Science and Engineering

28 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The non-stationary property of the streaming data often evinces the concept change, which
is either regular or random. Detection of such change points is termed as handling concept
Handling Concept drift, or concept drift detection. The contemporary models are proven to be considerable to
Drift by Distribution detect concept drift. However, majority of these models are dependents of labeled data,
57 Similarity Measures which is impractical in most conditions. In this paper, Handling Concept Drift by
in High Dimensional Distribution Similarity Measures (HCDDSM) in High Dimensional Data Stream is
Data Streams proposed. The HCDDSM is not depends on labels to denote the concept drift. The
experimental study portrayed that HCDDSM is scaled model to detect concept drift, which
is to detect concept drift with minimal delay, maximal drift detection rate.

Agriculture is the primary source of food production. Protecting agricultural activities

ensures a stable and reliable food supply for the growing global population. Without
adequate protection, crops can be affected by pests, diseases, extreme weather events, and
other factors that may lead to food shortages.
Agriculture is a field which need to be concentrated more to be equipped with latest and
modern technologies to improve the yield and protecting plants from harmful insects and
Automatic insect pests.
detection for The reduced crop yields will result in starvation and insufficient food supplies. So we need
sustainable coconut an efficient and effective method of detection and diagnosis are required to increase food
58
farming by leveraging productivity. Traditional methods are based on the human experts and manual. This will
IoT and Multi Model lead to time consuming. Human expertise may not be available as required in everywhere.
Deep Neural Networks Sometimes it is very difficult to large areas. To overcome this difficulty an automated
advanced technology needs to be adopted.
Methodology: Internet of Things (IoT) sensors can be effectively utilized to capture the real
time data. Different types of sensors can be used to capture the data like temperature,
humidity, soil moisture, and pH. These sensors need to be connected using sensor
networks. These sensors can also be equipped with disease-specific detectors, such as those
for pathogens or indicators of plant stress.

School of Computer Science and Engineering

29 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Computational diagnoses for knee joint disorders are gaining importance due to their
Knee Joint disorder essentiality. The existing modalities for sample extraction are static in nature which suffers
diagnosis using a in producing better outcomes. This research aims to focus an generating a new modality
59
novel modality and using the vibration and acoustic sensors for data collection and the deep learning approaches
deep learning models will be applied for accurate prediction of different knee joint disorder levels at their earlier
stages.

Infant mortality remains a significant public health challenge, particularly in low-resource

settings. Machine learning (ML) offers a promising approach to predicting infant mortality
by identifying critical risk factors and enabling early interventions. This study explores the
development and application of ML models to predict infant mortality, leveraging diverse
data sources including demographic information, maternal and neonatal health indicators,
socio-economic factors, and healthcare accessibility metrics. By employing supervised
learning techniques, such as logistic regression, decision trees, random forests, and neural
Infant mortality
networks, we aim to achieve high predictive accuracy and interpretability. The dataset
prediction in india
60 comprises historical health records and population surveys, which are preprocessed and
using machine
subjected to feature selection to enhance model performance. The models are evaluated
learning
using standard metrics like accuracy, precision, recall, and the area under the receiver
operating characteristic curve (AUC-ROC). Preliminary results indicate that ML models can
significantly improve the prediction of infant mortality compared to traditional statistical
methods. This research underscores the potential of ML in informing public health
strategies, optimizing resource allocation, and ultimately reducing infant mortality rates.
Future work will focus on refining models through larger datasets, real-time data
integration, and collaboration

School of Computer Science and Engineering

30 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This project aims to develop a personalized movie recommendation system using
collaborative filtering techniques. By leveraging user preferences and behavior, the system
predicts and suggests movies tailored to individual tastes. The collaborative filtering
Personalized Movie
approach identifies patterns and similarities between users and items, utilizing both user-
Recommendations
61 based and item-based methods to enhance recommendation accuracy. The project involves
Using Collaborative
data collection, preprocessing, and the implementation of collaborative filtering algorithms,
Filtering Technique
followed by evaluation and optimization to ensure high-quality recommendations. The
outcome is a user-centric recommendation system that provides relevant and personalized
movie suggestions, improving user satisfaction and engagement.

School of Computer Science and Engineering

31 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Vitamin D deficiency is a widespread health concern in India, affecting a significant portion
of the population and leading to various health complications. This study aims to predict
vitamin D deficiency using machine learning techniques, leveraging data on demographics,
lifestyle factors, dietary intake, and health indicators. By employing advanced algorithms
Predicting Vitamin D such as logistic regression, decision trees, random forests, and neural networks, we develop
Deficiency in India predictive models that accurately identify individuals at risk of vitamin D deficiency.
62
Using Machine Approach includes the collection and preprocessing of data from diverse sources, feature
Learning selection, model training, and evaluation to ensure robust predictions. The predictive models
not only achieve high accuracy but also provide insights into the key factors contributing to
vitamin D deficiency. This machine learning-based system can be utilized for early
identification and personalized intervention, aiding healthcare providers in addressing
vitamin D deficiency more effectively.

rain tumor diagnosis using MRI scans poses significant challenges due to the complex
nature of tumor appearances and variations. Traditional methods often require extensive
manual intervention and are prone to human error, leading to misdiagnosis and delayed
treatment. Current approaches primarily include manual examination by radiologists and
Refinement of neural conventional machine learning techniques. These methods rely heavily on feature extraction
network algorithms and classification algorithms, which may not capture the intricate patterns present in brain
63 for accurate brain MRI images. Conventional techniques often suffer from limited accuracy and
tumor classification generalizability, mainly due to the high variability in tumor appearance and the subjective
using MRI images nature of manual interpretation. Additionally, traditional machine learning models may
struggle with the high-dimensional data inherent in MRI images. To address these
limitations, our research introduces a deep learning-based model utilizing convolutional
neural networks (CNNs).Our model employs a sequential CNN architecture with multiple
convolutional, max-pooling, and dropout layers, followed by dense layers for classification.

School of Computer Science and Engineering

32 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The project titled "Sales Forecasting System for E-commerce Food Products" focuses on
developing an advanced analytical model to accurately predict sales in the rapidly growing
online food market. The project leverages historical sales data, coupled with customer
purchase patterns, seasonal variations, and market trends, to build a robust forecasting
system. Using a combination of time series analysis and machine learning algorithms, such
Sales Forecasting
as ARIMA and Random Forest, the system aims to provide precise sales predictions that
System for E-
64 help e-commerce platforms manage inventory more efficiently, optimize pricing strategies,
commerce food
and enhance overall customer satisfaction. The abstract would further detail the validation
Products
of the model against actual sales figures to ascertain its accuracy and reliability. The
outcomes of this project are anticipated to offer significant insights into demand forecasting,
facilitating better decision-making for e-commerce businesses specializing in food products.
This project not only enhances operational efficiencies but also contributes to the academic
body of knowledge in predictive analytics and e-commerce logistics.

We can collect the details about the daily activities of the user through various wearables
65 Diet plan prediction
and suggest suitable food based on the target set.

School of Computer Science and Engineering

33 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In today's educational landscape, understanding and predicting student performance is
crucial for enhancing academic success and personal growth. This project aims to leverage
data science and machine learning techniques to predict student performance based on
various academic, socio-economic, and demographic factors.

The project involves collecting a comprehensive dataset of student information,

preprocessing the data, and applying feature engineering to extract meaningful insights.
Predictive Analytics Various supervised learning algorithms, such as linear regression, decision trees, and
for Student support vector machines, will be employed to build predictive models. The performance of
66
Performance Using these models will be evaluated using metrics such as accuracy, precision, recall, and F1-
Machine Learning score.

Furthermore, the project will explore the impact of different features on student
performance, providing actionable insights for educators and policymakers to improve
teaching strategies and student support systems. By the end of the project, students will gain
hands-on experience in data preprocessing, model building, and evaluation, along with a
deeper understanding of how data science can be applied to real-world educational
challenges.

School of Computer Science and Engineering

34 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 As the world moves towards smarter and more efficient energy usage, integrating Internet of
Things (IoT) with machine learning for home energy management has become a promising
solution. This project aims to design and implement a Smart Home Energy Management
System (SHEMS) that leverages IoT devices and machine learning algorithms to monitor,
predict, and optimize energy consumption in real-time.

The project will start by setting up a network of IoT sensors and devices to collect data on
energy usage, environmental conditions (temperature, humidity, etc.), and appliance status
within a home. This data will be transmitted to a central server or cloud platform for storage
and analysis.

Smart Home Energy Machine learning models, including regression algorithms and time series forecasting
Management System methods, will be developed to predict future energy consumption patterns. These models
67
Using IoT and will be trained on the collected data to provide accurate and reliable predictions.
Machine Learning Additionally, optimization algorithms will be used to develop strategies for reducing energy
consumption during peak hours and improving overall energy efficiency.

The SHEMS will also feature a user-friendly interface that provides real-time monitoring
and control of home appliances, allowing users to make informed decisions about their
energy usage. Alerts and recommendations will be generated to help users minimize their
energy footprint.

By the end of this project, students will gain hands-on experience with IoT device
integration, data collection and preprocessing, machine learning model development, and
system optimization. They will also understand the practical applications of combining IoT
and machine learning to promote energy efficiency and sustainability in smart homes.

School of Computer Science and Engineering

35 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Managing diabetes requires careful monitoring of diet and health metrics to maintain blood
glucose levels within a healthy range. This project aims to develop a Diabetic Diet
Recommendation and Health Tracking System using data science techniques to provide
personalized dietary advice and track key health metrics for diabetic patients.

The project will start with the collection of a comprehensive dataset that includes nutritional
information of various foods, diabetic patients’ dietary preferences, medical history, and
health metrics such as blood glucose levels, HbA1c, and physical activity. Data
preprocessing steps will involve cleaning the data, handling missing values, encoding
categorical variables, and normalizing numerical features.

Machine learning algorithms, such as collaborative filtering and content-based filtering, will
Diabetic Diet be implemented to develop personalized diet recommendation models. These models will be
Recommendation and trained to suggest meal plans and food items that align with the patients’ dietary needs and
68 Health Tracking preferences, aiming to maintain stable blood glucose levels.
System Using Data
Science The health tracking component will include data collection from wearable devices or
manual input by patients, tracking metrics such as blood glucose levels, physical activity,
and medication adherence. Time series analysis and predictive modeling techniques will be
employed to analyze trends and predict potential fluctuations in blood glucose levels.

The system will feature a user-friendly mobile or web application where patients can input
their health metrics, receive dietary recommendations, and track their progress over time.
The application will provide real-time feedback and alerts to help patients manage their
diabetes effectively.

By the end of this project, students will gain hands-on experience in data preprocessing,
machine learning model development for recommendation systems, time series analysis, and
creating user-centric applications. They will also learn how data science can be applied to

School of Computer Science and Engineering

36 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The spreading of false information over the internet, highly impacts the society and causes
social and economic loss to the nation. One of the most important factors of the increase in
the number of misleading information is that all the users of any particular digital platform
usually do not verify the authenticity of the information. The primary objective of this
proposed work is to provide a solution to society in order to reduce the social and economic
losses caused by spreading false information circulated digitally. Block-chain is a
distributed, decentralized network that conducts peer-to-peer transactions, meaning that it
does not have a main or central entity to operate. Here we present a literature survey on
various approaches towards this problem statement and investigate a Blockchain-based
Detection of False
solution that focuses on all types of digital content by providing the end user to validate the
information in digital
69 information along with the evidence for data that was collected at each step.
media using Block
chain Technology

This project focuses on developing a system for automated detection and classification of
Crop Disease
diseases in crops using machine learning and image processing techniques. By analyzing
70 Detection and
images of diseased plants, the model will identify the type of disease and recommend
Classification
appropriate treatments, aiding farmers in crop management and yield optimization.

School of Computer Science and Engineering

37 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Certainly! Here's an elaboration of the abstract for the project "Water Management System
for Sustainable Agriculture: A Data-Driven Approach":

Title: "Water Management System for Sustainable Agriculture: A Data-Driven Approach"

Abstract:

Water scarcity is a significant challenge for agriculture, impacting crop yields and food
security. This project aims to address this challenge by developing a data-driven water
management system for sustainable agriculture. The system will integrate data from various
Water Management
sources, including soil moisture sensors, weather stations, crop models, and water
System for
availability databases.
71 Sustainable
Agriculture: A Data-
The collected data will be analyzed using data analytics and machine learning techniques to
Driven Approach
optimize irrigation scheduling. By considering factors such as soil type, crop type, weather
conditions, and water availability, the system will recommend precise irrigation schedules to
farmers. This will help reduce water wastage, improve crop yields, and conserve water
resources.

Furthermore, the project will develop a user-friendly interface for farmers to access
irrigation recommendations and monitor water usage. This will empower farmers to make
informed decisions about water management, ultimately promoting sustainable agriculture
practices. Overall, this project not only enhances the students' data science skills but also
contributes to addressing water scarcity and ensuring food security.

The main purpose is to develop machine learning algorithms for predicting the percentage
of Student Placement Prediction using machine learning based on the data related to the
Student Placement
university’s academic reputation, opportunities of the city where the university is located,
72 Prediction using AI |
facilities, and cultural opportunities of the university. It analyzes the previous year’s
Machine Learning
historical data and predicts placement of current students and aids to increase the placement
percentage of the institutions

School of Computer Science and Engineering

38 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Student organizations offer chances for skill development, community involvement, and
personal growth, making them an essential component of the university experience. But for
these clubs to make the biggest difference in campus life, we need to maximize their
support. This abstract presents a thorough plan for enhancing community liveliness, student
club management, and involvement through the usage of a suggested club hub application.
Club Hub-An
The proposed software combines physical locations, policies, and training to simplify club
application for
73 management operations. Its objective is to facilitate students ability to locate and join
connecting the clubs
groups. The application allows students to stay updated about upcoming club events,
with students.
encouraging increased participation and improving the collaborative component of club
membership. This helps to connect the like-minded people not only for club, it can be
connected with alumni so that students can interact with the alumni for career guidance and
other activities. As there is a gap in communication between event updates to the students, it
will be helpful for the students for the academic as well as career based activities.

The agriculture industry in India is deteriorating, which has an impression on the

ecosystem's output potential. There is an crucial necessity to fix the problem in the domain
in order to restore vitality and propel it forward. The learning suggestions an e-Agriculture
Application built on a framework that includes KM-Knowledge base and Monitoring
modules. Farmers require information throughout the farming cycle in order to make
effective decisions. The necessary information is dispersed across the system, and it
comprises real-time data such as market pricing and current production level statistics, as
well as primary agricultural knowledge. A knowledge dataflow model is built to connect
Field Monitoring and
multiple dispersed sources to crop structures. With the progress of technology, the world is
74 Automation Using
becoming more automated, with automated operations replacing manual procedures since
IoT Agriculture
they are more energy efficient and need less man labor. The study presents the benefits of
using ICT in the Indian agricultural industry, demonstrating the road for rural farmers to
replace some of the traditional approaches. Monitoring modules are demonstrated utilizing
various sensors, the inputs for which are provided from the Knowledge base. The system is
prototyped utilizing TI CC3200 Launchpad linked sensor modules and other essential
electrical equipment. A comparison of the created system and the current systems is
performed. The technology overcomes the limits of traditional agricultural techniques by
efficiently using water resources and lowering labor costs.

School of Computer Science and Engineering

39 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This paper proposes a comprehensive solution to the challenges faced in university
classrooms regarding energy wastage and damage to electrical infrastructure. In many
educational institutions, it is common for students to neglect turning off lights and fans after
class, leading to unnecessary energy consumption. Moreover, instances of vandalism, where
students damage electrical switchboards and fan regulators, contribute to maintenance costs
and disruptions in the learning environment.To address these issues, our research introduces
Smart Classroom a Smart Classroom Management System (SCMS) leveraging modern technologies. The
Management System: SCMS integrates various sensors, such as occupancy sensors and motion detectors, to
Addressing Energy monitor classroom activity in real-time. Additionally, it incorporates smart switches and
75 Conservation and regulators equipped with IoT capabilities, enabling remote control and monitoring of
Preventing Vandalism electrical devices.Through machine learning algorithms and predictive analytics, the SCMS
through Technology learns and adapts to patterns of classroom usage, optimizing energy usage by automatically
Integration" adjusting lighting and ventilation based on occupancy levels. Furthermore, the system
employs security features to detect and deter vandalism attempts, providing alerts to
administrators in case of suspicious activity.Our proposed solution aims to promote
sustainability by reducing energy wastage while also enhancing the safety and security of
classroom infrastructure. By implementing the Smart Classroom Management System,
universities can achieve significant cost savings, minimize environmental impact, and create
a more conducive learning environment for students and educators alike.

Internet of Things (IoT) is not only a promising research topic but also a blooming industrial
trend. Although the basic idea is to bring things or objects into the Internet, there are various
approaches, because an IoT system is highly application oriented. This paper presents a
wireless sensor network(WSN)-based IoT platform for wide area and heterogeneous sensing
Design of a Wireless
applications. The platform, consisting of one or multiple WSNs, gateways, a Web server,
Sensor Network-
and a database, provides a reliable connection between sensors at fields and the database on
Based IoT Platform
76 the Internet. The WSN is built based on the IEEE 802.15.4e time slotted channel hopping
for Wide Area and
protocol, because it has the benefits such as multi-hop transmission, collision-free
Heterogeneous
transmission, and high energy efficiency. In addition to the design of a customized hardware
Applications
for range extension, a new synchronization scheme and a burst transmission feature are also
presented to boost the network capacity and reduce the energy waste. As a result, the
proposed platform can fulfill the high throughput requirement for high-rate applications and
the requirement of long battery life for low-rate applications at the same time.

Enhancing E- With the increasing reliance on e-learning platforms, ensuring these systems are accessible
Learning to all users, including those with disabilities, is essential. This project aims to develop an e-
77 Accessibility through learning platform that leverages Artificial Intelligence (AI) and inclusive design principles
AI and Inclusive to enhance accessibility, providing a seamless and equitable learning experience for all
Design students.

School of Computer Science and Engineering

40 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Declarative problem solving technique that combines powerful performance solving
capabilities with an easy-to-use modeling language. Here, we create an logic-based method
A Tool for for one of the most extensively researched academic course time table issues: academic
Developing Academic course scheduling (ACTT). The resulting system, reads a standard input format's ACTT
78 Course Time Table instance and transforms information into a set of facts. These data are then coupled with a
using Logic first-order encoding for the purpose of solving ACTT, which may then be handled by any
Programming commercial logic programming system. We demonstrate the competitiveness of our method
by comparing it empirically to the best known bounds that have been attained thus far
through specialized implementations

Intelligent The goal of this initiative is to leverage IoT technologies to transform transportation
Transportation networks. Sensor-equipped connected cars may talk to infrastructure and other connected
Systems and cars to improve safety and traffic flow. Data analytics is used by intelligent transportation
79
Connected Vehicles systems to control traffic, increasing both overall efficiency and safety. Furthermore,
in the Transportation sensors are used by smart parking systems to expedite parking availability and shorten
Industry search times.

The main idea of this AI based Project is to create an automated defense and tackle system
consists of Automated Ranged gun, self activating landmines, landmine detecting and
AI Defence and
80 mapping system. This will help our army to gain a better command of defense in hostile
Tackle System
areas and borders and will help in reducing of lives. The Landmines has been a major issue
faced by both the army and civilians in some areas .

Collection of data of multiple modalities (understanding why to use the modality from
Student performance
literature survey) that aid in drawing feedback on the performance of students by studying
analysis using
81 the correlation of primary and derived features extracted to validate few hypotheses and
multimodal data
applying a learning methodology (ML, DL, or RL) for training and testing. This can help in
analysis
obtaining a feedback doesn't rely merely on the score obtained by the student.

The voting system proposed here is a complete rewrite of the traditional EVM model which
Blockchain Based uses fingerprint authentication and face recognition for verifying the user for credentials.
Secured Voting With cloud-based blockchain, an immutable storage system for transaction information,
System Using Face many cloud based applications which prioritize privacy and authenticity are shifting from
82
Recognition and traditional databases to blockchain implementations. The use of blockchain for storage of
Fingerprint polling results and the usage of fingerprint and face maps for authentication are the key
Authentication features of this implementation furthermore this project only acts as a proof of concept for
implementing a voting system using biometrics for authentication.

School of Computer Science and Engineering

41 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 There are a lot of tools and devices available to teach mathematics on a basic level like
identifying numbers and counting. For example, Math tool by Math Adventures. The
individual with autism requires personal attention to understand the mathematics concepts
step by step. The traditional learning method of Mathematics is not suitable for autistic
children. Educationalists identify the gaps and frame the different curricula and teaching
methods which differ from person to person.

AI-based Math tool This research work will produce effective teaching strategies and tools for autistic children.
83
for autistic children It needs the process to refine the tool based on observations and feedback. We can have all
teaching methods manually. To help the teachers or tutors an interactive Mathematics
learning tool is very supportive as well as for autistic children.

To make this research successful, we need to do a multidisciplinary collaboration like user-

centered design, participatory design, educational research, cognitive psychology, human-
computer interaction, assistive technology and mixed methods research.

Web scraping is a beautiful and helpful technique to extract raw data from a website. The
data analyst can use this extracted information to store in various formats and analyze for
Web Scraping Using
84 decision-making purposes. Beautiful Soup is a Python package for parsing HTML and XML
Beautiful Soup
documents and extracting data from them. It works as a favorite parser to yield idiomatic
searching, navigating, and modifying the parse tree.

School of Computer Science and Engineering

42 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Big data mining has emerged as an active field of interest, and traditional data mining
approaches frequently fail to handle the complexities associated with massive datasets. One
of the most extensively used strategies for big data classification is MapReduce, which
combines the map and reduce processes.
A new approach logistic regression based Extreme Learning Machine classifier-based
Logistic Regression MapReduce framework is proposed in this paper for big data classification. The IELM
and ELM Classifier algorithm is formed by integrating the ELM technique with Principal Component Analysis
85
for Big Data to overcome the multicollinear problem and enhance the training and testing time. The
Classification Tasks proposed algorithm will improve the optimal parameter selection of the IELM classifier,
which increases classification accuracy and reduces the error rate.
The performance of the proposed work will be evaluated using different performance
metrics such as Accuracy, Sensitivity, Specificity, Computational time, F1-Score, Mathews
correlation coefficient and scale-up factor. The proposed big data classifier will be effective
for both single-class and multi-class classification.

Continual learning, the ability of a model to learn continuously from a stream of data
without forgetting previous knowledge, remains a challenging task in machine learning. A
key obstacle is the efficient management of training data, particularly when the availability
of diverse samples is limited. Our aim for this work is to introduce a novel approach for
Sample generation for
sample generation aimed at enhancing continual learning by creating multiple new samples
86 efficient continual
from a single original sample, each incorporating features from multiple other samples. By
learning
leveraging advanced image blending techniques, we will synthesize diverse training samples
that retain the intrinsic properties of the original data while introducing variations that
reflect the characteristics of the entire dataset. This method not only augments the training
set but also improves the robustness and generalization capabilities of the learning model.

School of Computer Science and Engineering

43 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The University Online Facility Management System (UOFMS) with Hostel Integration is a
transformative digital platform designed to revolutionize the management of university
facilities and student accommodations. This project addresses the growing need for efficient
resource allocation, maintenance scheduling, and real-time monitoring capabilities within
academic institutions. UOFMS offers a comprehensive suite of features including user
University Online authentication, facility booking, maintenance management, asset tracking, hostel
Facility Management management, and real-time monitoring and reporting. By integrating hostel management
87 System (UOFMS) functionalities seamlessly with existing facility management systems, UOFMS aims to
with Hostel streamline operations, enhance user experiences, and optimize resource utilization across
Integration the university campus. Through meticulous planning, systematic design, rigorous
development, and comprehensive evaluation, this project endeavors to deliver a robust and
user-friendly solution that meets the diverse needs of university stakeholders. The
successful implementation of UOFMS holds the potential to significantly improve
operational efficiency, enhance campus infrastructure management, and elevate the overall
quality of university services.

In industrial settings, unplanned equipment failures can lead to significant downtime and
financial losses. This project aims to develop a predictive maintenance system using
machine learning techniques to forecast equipment failures before they occur. By analyzing
historical sensor data and maintenance records, the model will identify patterns and
Predictive
anomalies that precede equipment malfunctions. Techniques such as time series analysis,
Maintenance for
classification, and regression will be employed to create predictive algorithms. The project
88 Industrial Equipment
will also include the development of a user-friendly dashboard to visualize predictions and
Using Machine
provide actionable insights for maintenance scheduling. This system aims to improve
Learning
operational efficiency, reduce maintenance costs, and enhance the reliability of industrial
equipment.

INNOVATIVE
METHODOLGIES Improving the attention span of students can have a significant impact on their learning
89 FOR IMPROVISE outcomes. In this project we are planning to find and implement out several innovative
ATTENTION SPAN methodologies to increase the attention span of students in the class rooms.
OF STUDENTS

School of Computer Science and Engineering

44 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 An intrusion detection system (IDS) is a vital piece of cybersecurity infrastructure that
detects unauthorized access, misuse, or irregularities in a computer network or system. Its
major job is to monitor network and system activity in real- time, look for signals of
malicious behavior or policy violations, and then create warnings or take preventive
measures to minimize such threats. An IDS abstract often defines its essential objectives,
methodology, and significance in the field of cybersecurity. It underlines the need to quickly
detect and respond to security breaches to protect sensitive information, maintain system
Enhancing Network integrity, and assure continuous business operations. One of the biggest issues facing
Security Through advancements in network security is accurately identifying and stopping breaches. Using the
Advanced Intrusion Friday-WorkingHours-Afternoon- DDos.pcap_ISCX11 dataset, this study aims to determine
Detection: A Fusion which algorithm is more accurate in detecting intrusions, more especially Distributed Denial
90 of Dimensionality of Service (DDoS) attacks. The dataset was selected because it is relevant to real-world
Reduction and circumstances, particularly those involving Friday afternoon work hours. A comparison
Machine Learning analysis is carried out to find out how well different classification algorithms perform in
Classification for terms of intrusion detection. The project entails tedious performance evaluation, substantial
Improved Accuracy parameter tweaking, and algorithm experimentation. Finding the algorithm that best handles
the dataset's distinctive patterns is the goal; eventually, this will help intrusion detection
systems evolve. The goal of this study's findings is to offer helpful details about the best
intrusion detection system for the given dataset. Finding the most accurate algorithm is
crucial to improving network security as cyber threats keep getting more sophisticated.
Professionals in cybersecurity who wish to install strong intrusion detection systems may
find the findings helpful. They can also guide the creation of fresh algorithmic tactics that
will be required in the future to counter new dangers that surface.

At the moment, medical professionals use manual analysis of the patient's MR images to
pinpoint the location and size of the brain tumor. This is thought to be exceedingly time-
consuming and leads to an incorrect tumor detection. A tumor is an uncontrollably growing
brain tumor detection
clump of tissue. For the purpose of detecting brain tumors, we can employ the Deep
91 using deep learning
Learning architectures of VGG 16 (visual geometry group) transfer learning and CNN
techniques
(convolution neural network), sometimes referred to as NN (Neural Network). The model's
performance is to forecast whether a tumor will be seen in an image or not. Returns yes if
the tumor is present and no otherwise.

School of Computer Science and Engineering

45 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This project will attempts to throw light on the history of the almost extinct but once
AI Model for Lost widespread script of Kaithi. Originating in north India and becoming popular in north and
92 Languages in India: central parts of the country, the script was once widely used in order to maintain a variety of
"Kaithi" A Case Study records, and was popularly taught in schools. However, it was systematically removed from,
and subsequently lost in popular literary culture by the twentieth century.

School of Computer Science and Engineering

46 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Automated Financial
93 Automated Financial Trading System
Trading System

Today there are no certain methods by using which we can predict whether there will be
rainfall today or not. Even the meteorological department’s prediction fails sometimes. In
Rainfall Prediction this article, we will learn how to build a machine-learning model which can predict whether
94 using Machine there will be rainfall today or not based on some atmospheric factors. This problem is
Learning related to Rainfall Prediction using Machine Learning because machine learning models
tend to perform better on the previously known task which needed highly skilled individuals
to do so.

The integration of Internet of Things (IoT) technology and machine learning (ML)
algorithms has brought about significant transformations across numerous industries, with
"HarvestGuard:
agriculture being a notable beneficiary. Precision agriculture, geared towards enhancing
Optimizing Crop
95 crop yield, resource efficiency, and farm productivity, has emerged as a fertile ground for
Yield through IoT and
this collaborative endeavor. This review meticulously examines the latest progressions,
ML"
hurdles, and prospects in harnessing IoT devices and ML techniques within the realm of
agriculture.

Managing a dairy cattle farm involves several intricate tasks, including monitoring the
CLOUD
health of the animals, tracking milk production, managing breeding schedules, and
APPLICATION FOR
optimizing feeding regimens. A cloud-based application, Cattle Hub, can streamline these
96 DAIRY CATTLE
processes, enhance data accessibility, and improve overall efficiency. This document
MANAGEMENT
outlines the key features and benefits of Cattle Hub, a comprehensive dairy cattle
CATTLE HUB
management solution.

Abstract :This project focuses on implementing an IoT-based solution for smart agriculture,
IoT-Based Smart integrating RFID technology for animal tracking. RFID tags will be attached to livestock,
Agriculture with enabling farmers to monitor their location, health, and behavior in real-time. Additionally,
97
RFID Animal environmental sensors will collect data on factors such as temperature, humidity, and soil
Tracking: moisture, providing insights for optimized farm management practices and improved animal
welfare.

School of Computer Science and Engineering

47 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Context-centric proximity detection is a promising approach to defend against relay attacks
in many mobile authentication systems. Prior work demonstrated the effectiveness of a
variety of contextual sensor modalities for this purpose, including audio-radio environment
(ambient audio, WiFi, Bluetooth and GPS) and physical environment (temperature,
humidity, gas and altitude). We present a systematic assessment of co-presence detection in
the presence of a strong, context-manipulating attacker against unattended terminals. First,
we show that it is feasible to manipulate, consistently control and stabilize the readings of
Sensor-based
different acoustic and physical environment sensors using low-cost, off-the-shelf equipment.
Proximity Detection
98 Second, based on these capabilities and the strengthened threat model, we show that an
in the Face of Active
attacker who can manipulate the context gains a significant advantage in defeating context-
Adversaries
based co-presence detection. For systems that use multiple sensors, we investigate two
approaches based on machine learning techniques - features-fusion and decisions-fusion,
and show that both are vulnerable to contextual attacks but the latter approach can be more
resistant in some cases. We further consider other defensive approaches that may be used to
reduce the impact of even such a strong context-manipulating attacker. Our work represents
the first concrete step towards analyzing, extending and systematizing prior work on
contextual co-presence detection under a stronger, but realistic adversarial model.

Data is transformed into a hash code or cipher through the process of encryption in order to
prevent adversaries' illegal access. Secure keys and validation procedures would grant
Text Encryption
access to the content to the authorized user. You can begin with a text encryption cyber
99 Using Cryptographic
security project as a novice. You might learn how to dissect the architecture of algorithms
Algorithms
such as Caesar Cipher, Vigenere Cipher, Railfence Cipher, Autokey Cipher, Playfair
Cipher, Beaufort Cipher, etc.

Large-scale commercial, industrial and financial operations are becoming ever more
interdependent, and ever more dependent on IT. At the same time, the rapidly growing
interconnectivity of IT systems, and the convergence of their technology towards industry-
standard hardware and software components and sub-systems, renders these IT systems
increasingly vulnerable to malicious attack. This paper is aimed particularly at readers
concerned with major systems employed in medium to large commercial or industrial
Securing IT Systems
enterprises. It examines the nature and significance of the various potential attacks, and
100 from Cyber Crime in
surveys the defence options available. It concludes that IT owners need to think of the
IT Organization
threat in more global terms, and to give a new focus and priority to their defence. Prompt
action can ensure a major improvement in IT resilience at a modest marginal cost, both in
terms of finance and in terms of normal IT operation.

School of Computer Science and Engineering

48 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 School of Computer Science and Engineering
49 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 he rapid advancement of technology and the growing demand for continuous learning have
underscored the importance of skill development in today's dynamic job market. This
project presents the design and implementation of an "E-Learning Platform for Skill
Development," a comprehensive application aimed at facilitating lifelong learning and
professional growth.

The platform leverages interactive courses, tutorials, and assessments to provide users with
an engaging and effective learning experience. Key features include personalized learning
paths, real-time progress tracking, and gamification elements to enhance user motivation
and retention. Additionally, the platform incorporates a social learning component, allowing
users to interact, collaborate, and share knowledge with peers and experts.
E-Learning Platform
101
for Skill Development
Utilizing modern technologies such as cloud computing, artificial intelligence, and
responsive design, the platform ensures scalability, accessibility, and a seamless user
experience across various devices. The application supports a wide range of skills, from
technical domains like programming and data science to soft skills such as communication
and leadership.

This project aims to bridge the gap between traditional education and the evolving needs of
the workforce by providing a flexible, user-centric solution for skill acquisition and
development. The "E-Learning Platform for Skill Development" not only empowers
individuals to achieve their career goals but also contributes to the broader objective of
creating a skilled and adaptable workforce in an ever-changing global economy.

School of Computer Science and Engineering

50 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Evolutionary Algorithms form the base for creating Artificial Intelligence applications and
systems. Evolutionary Computation forms the basis of those algorithms that are used in day
to day and future applications. Computer networks as we know, form a major contributor to
those data sources. Big Data and all types of computer related data are found on the internet
and the internet forms sources of data all over the world. Therefore, data modulation
techniques form a basis of data transfer over the internet all over the world.
Therefore, Evolutionary Networks are the type of computer networks that evolve over
time due to the evolutionary nature of the algorithms involved. Our intention in the
paper is to create a fault free data modulation technique used in computer networking.
In the network, the timing is based on the evolutionary data of the customer. The
internet history of the customer is taken into consideration and the data modulation is
based on the customer history. How this is done is based on the pheromone
NETWORK concentration of ant colony optimization and the probable path of the bee algorithm.
OPTIMIZATION These evolutionary algorithms are fed to the firewall and routing tables of the network
102
USING GENETIC to form the evolutionary network generation and user behaviour. When the algorithms
PROGRAMMING. working behind firewall and routing tables are replaced by Evolutionary Algorithms,
then they start to showcase evolutionary behaviour. This is the foundation principle on
which this paper is based. Once the firewall and the routing table are evolved, the
computer network to which they are connected becomes Evolutionary Networks.

School of Computer Science and Engineering

51 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Preliminary performance assessment of an affordable real-time High-Sensitivity (HS)
Global Navigation Satellite System (GNSS) receiver. Specifically tailored to capture and
track weak Galileo E1b/c signals, this receiver aims to support research endeavors focused
on advancing GNSS signal processing algorithms, particularly in scenarios characterized by
pronounced signal attenuation. Leveraging System-on-Chip Field-Programmable Gate Array
(SoC-FPGA) technology, this design merges the adaptability of Software Defined Radio
Implementation of a (SDR) concepts with the the robust hardware processing capabilities of FPGAs. This
High-Sensitivity innovative approach enhances power efficiency compared to conventional designs relying
103 Global Navigation on general-purpose processors, thereby facilitating the development of embedded software-
Satellite System defined receivers. Within this architecture, we implemented a modular GNSS baseband
Receiver . processing engine, offering a versatile platform for the integration of novel algorithms. The
proposed receiver undergoes testing with live signals, showcasing its capability to process
GNSS signals even in challenging scenarios with a carrier-to-noise density ratio (C/N0) as
low as 20 dB-Hz, while delivering navigation solutions. This work contributes to the
advancement of low-cost, high-sensitivity GNSS receivers, providing a valuable tool for
researchers engaged in the development, testing, and validation of experimental GNSS
signal processing techniques.

School of Computer Science and Engineering

52 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Generally, the tenders or contracts are used by governments and companies to procure
goods and services. Wrongful tender management leads to huge loss in case of faulty
practices. This includes favoring of contractors, improper record maintenance, lack of
Smart Tender- transparency, hacking, data modification and other issues. To overcome this problem, we
Contract Management have used a simple and secure block chain technology and to secure by encryption coupled
104
System Using with indisputable block based architecture for transaction management. In this case we
Blockchain make use of block chain technology to secure transaction based documents along with
transactions such as tender documents, applications, bid proposals, company profiles, past
records, approving officer details, rejection details to ensure a completely transparent
tendering process.

Even if you are not a Botany student, you will have fun realizing how the leaves are
because their volume, prevalence, and unique characteristics can serve as an effective
measure to identify plant species. Explore the source code link to learn about this project's
Plant species
implementation from scratch. You will enjoy getting to know about the methods that include
105 identification n
image-based features. And, as you may have guessed already, this would be a machine
prediction using ML
learning classification project so you will be introduced to the implementation of
classification machine learning algorithms in great depth. You will also learn to benchmark
the significance of different classifiers in image classification problems.

This project focuses on using machine learning to detect phishing emails and enhance email
security. By analysing large datasets of emails, the system learns to identify patterns and
AI-Driven Phishing
characteristics typical of phishing attempts. Features such as email content, sender
Detection for
106 information, and link destinations are examined. The machine learning model, once trained,
Enhanced Email
can automatically flag and filter out suspicious emails before they reach the user's inbox.
Security
This approach helps in reducing the risk of phishing attacks, protecting sensitive
information, and improving overall cyber security.

School of Computer Science and Engineering

53 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 # Project Title Project Abstract
With the rise of digitalization, the security and integrity of replicated data have become
paramount concerns for organizations across various sectors. Blockchain technology offers a
groundbreaking solution to these challenges by providing a secure, transparent, and
immutable platform for data replication. This abstract explores the application of blockchain
technology to safeguard replicated data from unauthorized access, tampering, and data loss.

Blockchain, as a distributed ledger technology, ensures the integrity and immutability of

replicated data through its decentralized consensus mechanism. Each transaction is
cryptographically linked to the previous one, forming an unalterable chain of blocks. This
tamper-resistant nature of blockchain makes it inherently robust against unauthorized
Blockchain modifications or tampering attempts.
Technology for
Securing Moreover, blockchain employs advanced cryptographic techniques, including hashing and
Replicated Data: digital signatures, to secure replicated data against unauthorized access. Access controls and
107 Mitigating permissioned blockchain networks further enhance security by restricting data access to
Unauthorized authorized participants only. Additionally, blockchain's decentralized architecture eliminates
Access, single points of failure, mitigating the risk of data loss due to hardware failures or malicious
Tampering, and attacks.
Data Loss
Furthermore, the transparency and auditability offered by blockchain enable real-time
monitoring and tracking of data transactions. Any unauthorized access or tampering attempts
are immediately detectable, allowing for prompt intervention and remediation. This
transparency also fosters trust among stakeholders by providing a verifiable record of data
provenance and integrity.

In conclusion, blockchain technology presents a robust solution for protecting replicated data
from unauthorized access, tampering, and data loss. By leveraging its decentralized,
immutable, and transparent features, organizations can establish a secure and resilient
framework for data replication, ensuring the confidentiality, integrity, and availability of their

School of Computer Science and Engineering

54 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 India's population is steadily increasing, particularly in urban areas or cities. Population
growth poses a significant challenge, particularly in urban areas worldwide, including India,
because it pressures the availability of water, electricity, infrastructure, traffic, health care,
education, etc. Population growth is a significant challenge to India Vision 2040, which is
focused on India's sustainable development and improving the quality of life of its residents.
Decision-support systems are critical in any planning of how population growth affects
resources. A satellite-based decision-support system is highly sophisticated and detailed. City
planning stands as a cornerstone in shaping the trajectory of urban development. Without
careful consideration and strategic foresight, cities risk falling prey to haphazard growth and
Empowering India mismanagement. Unplanned cities often yield dire consequences, ranging from housing
Vision 2040: A shortages and traffic congestion to environmental degradation and social inequalities. These
Map-Based outcomes underscore the urgent need for a systematic approach to urban development. The
Digital Twin primary objective of the proposed work is to facilitate the creation of prosperous cities in
Systems to Predict India that thrive in the face of dynamic challenges and ensure the well-being and prosperity
108 Population of residents.
Growth Impacts In the proposed work, we offer a systematic and comprehensive approach to city planning
on India's i.e., decision support systems using a map-based digital twin system to predict population
Resources, growth impacts on India's resources, sectors, and economy. The proposed system aimed to
Sectors, and mitigate unplanned urbanization's adverse effects and foster sustainable growth and
Economy. development.
By integrating advanced methodologies and leveraging innovative technologies, our approach
empowers city planners and policymakers with the tools to steer urban development toward a
sustainable, prosperous, and equitable future.
The anticipated outcome of the proposed work is to empower India Vision 2040 by providing
a decision support system to predict and visualize population growth impacts on India's
resources, sectors, and economy for policymakers in the form of a map-based digital twin.
The proposed system will help eradicate slums, prevent economic recessions, maintain a high
standard of living, and preserve happiness indices even in the most extreme circumstances.
Through the proposed digital twin system, cities in India can transcend the limitations of

School of Computer Science and Engineering

55 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Abstract:
This research survey delves into recent publications and future enhancements in natural
disaster prediction and response, emphasizing the role of Artificial Intelligence (AI) in early
warning systems and disaster relief efforts. Through an exhaustive review of cutting-edge
literature and emerging trends, this study identifies key algorithmic solutions and innovative
strategies poised to revolutionize disaster management practices in the coming years.
Recent publications underscore the importance of rigorous test analysis based on various
parameters to evaluate the effectiveness of AI-driven early warning systems and disaster
response mechanisms. Parameters such as prediction accuracy, lead time, false alarm rate,
scalability, and real-time responsiveness are critical metrics in assessing the performance of
algorithmic solutions in mitigating the impact of natural disasters.
Natural Disaster
Among the diverse array of AI algorithms investigated, Machine Learning (ML) techniques,
Prediction and
particularly ensemble learning methods and deep learning models, emerge as promising
Response:
solutions for natural disaster prediction and response. Ensemble learning techniques, such as
Leveraging AI for
109 Random Forests and Gradient Boosting Machines, leverage the collective wisdom of multiple
Early Warning
models to improve prediction accuracy and robustness, while deep learning models, such as
Systems and
Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), excel in
Disaster Relief
capturing complex patterns from multi-modal data sources, including satellite imagery,
Efforts
weather forecasts, and sensor data.
Furthermore, recent research highlights the potential of Reinforcement Learning (RL)
algorithms in optimizing disaster response strategies and resource allocation decisions. RL-
based models can learn optimal policies for coordinating emergency response efforts,
evacuating affected populations, and distributing relief supplies in a timely and cost-effective
manner, thereby minimizing casualties and mitigating the socio-economic impact of disasters.
In addition to predictive analytics, recent publications emphasize the importance of anomaly
detection algorithms for identifying early warning signs of impending disasters. Advanced
anomaly detection techniques, such as Isolation Forests and Auto encoder-based approaches,
enable early detection of abnormal patterns in environmental data, such as seismic activity,
atmospheric pressure, and river water levels, enabling proactive intervention and disaster

It is quite difficult to manage the objectives of PCL being interdisciplinary involving students
ERP for the
and mentors from multiple disciplines. There will be the issues of ego clashes as well as
110 management of
maintaining a central evaluation platform. Though the PCL is a great idea, still management
PCL
of the components of PCL is a challenging task involving diversity in its nature.

School of Computer Science and Engineering

56 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Computer vision-based sign language recognition (SLR) systems have emerged as promising
tools for bridging the communication gap between hearing-impaired individuals and the
hearing world. This project proposes an innovative SLR system that leverages the strengths
of two powerful algorithms, YOLO (You Only Look Once) and LLM (Large Language
Models), to achieve real-time and accurate sign language translation. The proposed system
utilizes YOLO, a state-of-the-art object detection algorithm, to efficiently identify and
localize individual sign gestures in real-time image frames. YOLO's high-performance object
detection capabilities enable the system to handle complex backgrounds and multiple signers
simultaneously. Once the sign gestures are identified, the system extracts relevant features
Enabling Real-
from each gesture using image processing techniques. These features are then fed into an
Time
LLM network, a recurrent neural network architecture specifically designed for sequential
Communication:
111 data processing. LLM's ability to capture long-term dependencies makes it well-suited for
A Computer
recognizing sign language, which involves a sequence of hand gestures and facial
Vision Approach
expressions. The LLM network processes the extracted features and generates a sequence of
for ASL
sign language labels, representing the interpreted sign language phrase. This project offers a
cutting-edge computer vision solution designed to help people who are hard of hearing. By
utilizing sophisticated methods in image processing and machine learning, the system is able
to identify and identify a wide range of typical ambient noises and transform them into visual
representations. Following that, a Flask online application that has been easily connected
with Large Language Models (LLMs) processes these visual clues. The software improves
accessibility for individuals with hearing impairments by providing real-time text-based
descriptions of detected sounds through the integration of LLMs.

In today's digital world, mobile phones have become a constant source of distraction for
students, often hindering their focus and academic performance. To tackle this issue, we
present an innovative solution that combines real-time distraction detection with personalized
intervention strategies. Our system's core component is an advanced app usage tracking tool
that keepstabs on how pupils use their smartphones. Our technology gathers important
Real Time Mobile insights into individual digital behaviour patterns by evaluating data such as the frequency
Distraction and length of app usage, and classifying apps into different categories (e.g., educational,
Detection and social media, entertainment). Leveraging advanced machine learning algorithms, our system
112
Personalized processes this data to identify potential distraction patterns during study sessions. Through
Intervention for continuous analysis, the system can accurately predict the likelihood of a student becoming
Student distracted, enabling proactive intervention measures. Our system's capacity to offer
individualized interventions based on each student's particular learning preferences and style
is what makes it truly innovative. Our technology adjusts its intervention tactics according to
the needs and behavioral tendencies of each individual, unlike traditional one-size-fits-all
approaches.

School of Computer Science and Engineering

57 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 School of Computer Science and Engineering
58 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Monitoring the ration food materials in public distribution system (PDS) warehouses in India
is critical to ensure transparency, prevent leakages, and efficiently manage the distribution of
essential commodities to beneficiaries. Monitoring of ration goods in public distribution
warehouses in India faces several challenges, which can lead to inefficiencies, leakages, and
corruption. Corruption at various levels of the supply chain, including warehouse staff,
Monitoring of transporters, and local officials, can result in diversion of ration goods for illegal sale in the
Ration food black market. Limited transparency in warehouse operations makes it difficult to track the
materials in public movement of ration goods accurately. Inadequate documentation and record-keeping
113
distribution practices contribute to this problem. The objective of this work is to Implementing
System warehouse computerized inventory management systems allows for real-time tracking of stock levels,
in India inward and outward movements of goods, and expiry dates. Barcode or RFID tagging can be
used to uniquely identify each item and facilitate automated inventory tracking. A cloud
based mobile application is designed with Equipping field staff with mobile applications for
data collection and reporting enables them to record stock movements, conduct inspections,
and report any irregularities directly from the field. This promotes transparency and
accountability in the monitoring process.

Identification, traceability, and real-time tracking of commodities in the Supply Chain (SC)
have long been hampered by the heterogeneity of platforms and technologies employed. To
assure the continuous supply of commodities to the Fair Price Shops (FPS), proficient supply
chain management is essential. In this regard, decision-makers must organize a number of
tasks related to the acquisition and transportation of material resources, product delivery, and
Employing
the delivery of commodities at the proper FPS and at a suitable time. The main objective of
Machine Learning
this research is to employ Machine learning (ML) techniques, including a combination of
for Successful
deep analytics, IoT and real-time monitoring that can be used to improve visibility of Public
Automation in the
Distribution System (PDS) SC, thus helping customer experience and achieve faster delivery
114 Public
commitments. ML models and workflows do this by analyzing historical data from varied
Distribution
sources followed by discovering interconnections between the processes along the supply
System's Supply
value chain. Historical data can precisely predict the lead times and reduce any errors. ML
Chain and
algorithms are used to optimize transport routes and schedules, track goods during transit and
Operations
also to analyze real-time traffic data to determine the most efficient delivery routes. This
approach aids in identifying the most efficient and fastest driver’s stop sequence while
minimizing driving time, distance and reducing delays. Such features also help to monitor and
predict traffic patterns impacting delivery times, such as peak hours at logistics hubs. By
employing ML in the SC, the PDS can anticipate and proactively mitigate challenges.

School of Computer Science and Engineering

59 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The abnormalities in brain cells are the main causes for forming lesions in brain. These
abnormal lesions in brain lead to the formation of tumors in brain. Magnetic Resonance
Imaging (MRI) and Computed Tomography (CT) are the two different brain image scanning
methods. In this project work, MR images are used to scan the brain internal regions. Benign
and Malignant are the type of abnormal lesions in brain in which, benign can be treated by
radiation methods; whereas malignant lesions are treated through proper surgery by expert
radiologist.
Tumor is defined as an uncontrolled growth of cancerous cells in any part of the body.
Tumors are of different types and possess diverse characteristics and require different
treatments. At present, brain tumors are classified into primary brain tumors and metastatic or
malignant brain tumors. The primary tumors begin in the brain and are inclined to stay in the
Prediction Of brain; the metastatic or malignant tumors begin as a cancer elsewhere in the body and then
Stroke And Brain start to spread into the brain region. Due to the large amount of brain tumor images that are
Tumor Using currently being generated in the clinics, it is not possible for physicians to manually annotate
115 Computer and segment these images in a practical time. Hence, the automatic tumor detection and
Assisted segmentation technique has become inevitable. In conventional methods, brain tumors are
Diagnosis detected and diagnosed manually by expert radiologist. It is time consuming and error probe
Approaches process. Hence, it is not suitable for high population developing countries. Therefore, a
computer aided automatic brain tumor detection and diagnosis methods are preferred.
The proposed Heuristic histogram equalization technique enhances the abnormal regions in
brain image which is further used to segment the tumor region. In order to improve the
classification accuracy, the features are extracted from the enhanced brain image for tumor
detection using Local Ternary Pattern (LTP) features and Grey Level Co- occurrence Matrix
(GLCM). The performance of the classification is determined by classification accuracy
which is defined as the ratio between number of images correctly classified to the total
number of images. In this project work, the performance of the brain image classification is
improved using Co-Active Adaptive Neuro Fuzzy Inference System (CANFIS) classifier.
Further, the Normalized Graph cut segmentation algorithm is used to segment the tumor
region in the classified Glioma brain image, the performance of the tumor segmentation

Early diagnosis and treatment of thyroid problems, ranging from hyperthyroidism to thyroid
IoT-Powered
BGCNet: Thyroid cancer, depend heavily on the identification of thyroid illness. Utilising an actual dataset from
Disease Detection the University of California, Irvine (UCI), this work offers a thorough method for identifying
116 Enhancement in thyroid disease (TD). Preparing the data is essential to guaranteeing the dataset's quality. The
Healthcare with dataset is improved to make it more appropriate for deep learning (DL) applications by
IBESA based managing outliers, and normalizing data. The Improved Bald Eagle Search Algorithm
Feature Selection (IBESA), a cutting-edge optimization method inspired by bald eagle hunting behaviour, is
used in the feature selection phase.

School of Computer Science and Engineering

60 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In the intricate rhythm of the health of humans, where every beat of the heart holds the key to
life's prognosis comes an effort for predictive precision that drives innovation at the
combination of nature-inspired optimization and cutting-edge deep learning and this quest is
at the heart of the human health sector. The challenge lies in accurately predicting heart
disease, a pivotal factor in human life prognosis, amidst the complexity of physiological data.
Despite developments in Deep Learning (DL), a major obstacle in contemporary healthcare is
Optimizing Heart
the precise prediction of cardiovascular illness, an essential component of human life
Illness Prediction
prognosis. A novel way to improving the accuracy of heart disease predictions is proposed in
with Bacterial
this research. This approach involves mixing nature-inspired optimization methodologies
117 Foraging and
with principles integrated with deep learning concepts. In particular, the framework of
LSTM-RNN: A
Bacterial Foraging Optimization (BFO) integrated with Long Short-Term Memory based
Deep Learning
Recurrent Neural Networks (LSTM-RNN) is presented in order to effectively extract vital
Fusion
features from cardiac datasets via a combined approach. A comprehensive pattern analysis
using ADN and temporal analysis using LSTM-RNN are the next phases in the proposed
methodology, which attempts to improve prediction accuracy. These steps are followed by
pre-processing procedures and feature selection that is aided by BFO integrated to create a
synergy that enables the detection of key relationships within the data. This, in turn, leads to
increased predictive performance for binary and multivariate classification challenges.

QR codes is simple and efficient tool used in smartphones to obtain accurate indoor user
location. As an increasing number of geo-location services are exploiting the capabilities of
smartphones, most of which incorporate GPS location. The smartphones compasses to direct
Navigation for the user to the destination. With the help of accelerometers, it will be possible to estimate the
Indoor Location walking distances as an auxiliary information. An, idea in order to calculate the user position
118 Based On QR inside a building by using QR codes and Google maps. Navigation is the process or activity
Codes and Google of accurately ascertaining one’s position, planning and following the route to goal location.
Maps Mobile phones are devices merely used to communicate. Based on new techniques like GPS
and sensors, compass and accelerometers that can determine the
orientation of devices, location based applications coupled with augmented reality views are
possible.

School of Computer Science and Engineering

61 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Crude Palm Oil (CPO) is one of the most consumed edible oils in the country compared

to other major agro-commodity oils. Due to certain global factors in 2021, CPO prices in
India have increased more than twice. The purpose of the given study is to investigate the
Impact of Crude
daily price movements of Crude Palm oil prices in India based on the degree of movements of
Oil prices and
external macroeconomic factors such as Brent crude oil price and dollar-based daily
currency
exchange rates. The co- relationship between the above variables is studied from the
exchange rate
119 multivariate GARCH(CCC(Constant conditional correlation) and DCC(Dynamic Conditional
fluctuations on
correlational) GARCH) approach. The results indicate no instantaneous causing effects and
palm oil prices in
hence any dependency of CPO price on the currency exchange rate. However, there exists
India: Dynamic
some direct impact of global crude oil prices on the Indian CPO price fluctuations. Also, the
GARCH approach
degree of covariance is quite significant among the three currencies. This study would help
industry stakeholders and practitioners hedge against imminent market risks arising from
unanticipated changes in India's palm oil prices.

JAIN UMS - TD- A realtime project in which students will develop a TD-PCL tracker with student group
120
PCL tracker formation, allocation of supervisor, project journal management etc.

Over the past two decades one of the most leading disease in today's era is the lung cancer. It
is very crucial in the medical field to understand the through genomics of this cancer and
provide the computer based solutions to the doctors for further well diagnosis of the patients
Early detection of suffering from this type of cancer. The significant role of AI engineers comes into picture to
121 lung cancer using early detect the type of lung cancer the patent is about to suffer and insist the medical
GAN supervisors how to proceed further. In this investigation, the AI engineers will use various
ML models and GAN for early detection of the type of lung cancer. The research in this
project will investigate the performance of basic and advanced models employed for the lung
cancer early detection.

School of Computer Science and Engineering

62 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Distributed Denial of Service (DDoS) attacks pose a significant threat to the availability and
reliability of online services, leading to substantial financial losses and operational
disruptions. Traditional defense mechanisms often struggle to adapt to the evolving nature of
these attacks. This paper presents a novel approach for predicting and mitigating DDoS
attacks using advanced machine learning algorithms, offering a proactive and adaptive
defense strategy. The proposed method leverages a combination of supervised and
unsupervised learning techniques to identify and mitigate potential DDoS threats. Initially,
historical network traffic data is used to train a supervised learning model, which can predict
the likelihood of an impending DDoS attack based on identified patterns and anomalies.
Feature extraction techniques, such as Principal Component Analysis (PCA) and statistical
A novel approach analysis, are employed to enhance the accuracy and efficiency of the model. In parallel, an
for predicting and unsupervised learning algorithm is deployed to continuously monitor real-time network
mitigating DDoS traffic, detecting abnormal behaviors that may indicate a DDoS attack. This dual approach
122
attacks using ensures both preemptive predictions and immediate detection of ongoing attacks. Once an
Machine learning attack is predicted or detected, the system automatically activates mitigation protocols,
algorithms including traffic filtering and rate limiting, to minimize the impact on network resources.
Comprehensive experiments and evaluations demonstrate that our machine learning-based
approach significantly improves the detection accuracy and response time compared to
traditional DDoS mitigation techniques. The results indicate that integrating predictive
modeling with real-time anomaly detection provides a robust framework for defending
against DDoS attacks, reducing false positives and enhancing overall network security. In
conclusion, this paper introduces an innovative machine learning-driven methodology for
predicting and mitigating DDoS attacks, combining predictive analytics with real-time
detection to offer a dynamic and effective defense mechanism. This approach represents a
substantial advancement in cybersecurity, with the potential for broad application in
protecting critical online infrastructure from the growing threat of DDoS attacks.

Amyotrophic Lateral Sclerosis (ALS) is a progressive neurodegenerative disease that affects

nerve cells in the brain and spinal cord, leading to muscle weakness and paralysis. Early
Machine Learning- detection is crucial for managing and potentially slowing disease progression. This project
Based Early aims to develop a machine learning-based system for the early detection of ALS using
Detection of magnetic resonance imaging (MRI). By analyzing MRI data, the system will leverage
123
Amyotrophic advanced algorithms to identify biomarkers indicative of ALS. The proposed solution will
Lateral Sclerosis involve the extraction of relevant features from MRI scans and the application of supervised
Using MRI learning techniques to classify patients at risk of ALS. The objective is to provide a non-
invasive, accurate diagnostic tool to aid neurologists in early intervention, improving patient
care and outcomes.

School of Computer Science and Engineering

63 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Emergency medical services in urban areas often face challenges such as traffic congestion,
limited resources, and unpredictable demand patterns, leading to delays in response times.
This project proposes to develop a smart ambulance system that leverages game theory
principles to optimize ambulance dispatch, routing, and resource allocation. Game theory
provides a powerful framework for modeling strategic decision-making situations with
multiple players and conflicting interests. The research methodology involves formulating the
Design of a Smart
ambulance dispatch problem as a game-theoretic model, defining players like ambulances and
Ambulance
124 hospitals, strategies, payoffs, and equilibrium conditions. Game theory algorithms, such as
System using
non-cooperative or cooperative games, will be developed to make optimal decisions for
Game Theory
ambulance routing, resource allocation, and hospital selection while considering real-time
data on traffic, hospital capacities, and demand. The proposed system will be implemented
using simulations and evaluated against traditional approaches based on metrics like response
times and resource utilization. Expected outcomes include novel game theory models,
efficient dispatch algorithms, performance benchmarking, insights into real-world
applications, and academic contributions to game theory for emergency response optimization.

School of Computer Science and Engineering

64 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Addressing the pivotal challenge of discerning deepfake alterations specifically within the
realm of human faces, this study delves into the nuanced task of identifying such
manipulations in the context of films. The advent of sophisticated algorithms such as
Deepfake and Face2Face has rendered conventional image forensics approaches inadequate
for video footage because of compression issues. Using deep learning techniques, the
suggested method presents two unique networks, Meso-4 and MesoInception-4, that are
intended to effectively detect hyperrealistic forgeries. Since standard approaches in video
forensics have limitations, the focus is on mesoscopic picture features. A formidable face-
swapping method called "deep fake" is being closely examined for its abuse in the production
of misleading information. By training two autoencoders, EA and EB, in tandem, the
technique allows the target person's face to be seamlessly replaced with another person's.
Although Deepfake has become more and more popular, the article points out some of its
Development a
flaws, including difficulties in handling facial occlusions. We also look at the Face2Face
Trusted Model for
125 methodology, which is a real-time facial reenactment method. Using pre-recorded films of the
Facial Video
target, this non-deep learning method overlays the target's face with a modified blend shape
Forgery Detection.
that mimics the source's facial emotions. To address these issues, the study presents Meso-4
and MesoInception-4, two low-layer systems that concentrate on mesoscopic image
characteristics. A recently produced Deepfake video dataset is one of the datasets used to
assess the networks. The results show remarkable detection rates that are higher than 98% for
Deepfake and 95% for Face2Face. The study clarifies the significance of interpretability for
deep neural networks and offers insights into how these networks recognize fake material.
Through comprehending the behavior of the network, the study advances the field of deep
learning-based video forensics. The suggested technique heavily relies on terms like
autoencoders, mesoscopic picture characteristics, and deep learning, emphasizing the
importance of these ideas in the fight against the spread of misleading video material.

School of Computer Science and Engineering

65 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The electrification of urban transportation has brought about a transformative shift in
mobility solutions, with two-wheeler electric scooters gaining popularity for their eco-
friendliness and efficiency. This paper delves into the application of Sinusoidal Pulse Width
Modulation (SPWM) inverters to control five-phase induction motors in two-wheeler electric
scooters. We provide an in-depth examination of the theoretical foundations, practical
implementations, advantages, and challenges associated with this technology. Additionally,
we discuss its implications for achieving sustainable urban mobility and the potential for
Sinusoidal PWM further advancements in this field. The key focus of this research is the implementation of
Inverter Fed Fault sinusoidal PWM inverters to achieve precise control of the motor, resulting in smoother
Tolerant Five- operation, reduced torque ripple, and improved overall performance. The control strategy
126 Phase Induction leverages the inherent advantages of a five-phase motor, such as increased torque density and
Motor Control for fault tolerance.
Two-Wheeler The study also addresses the challenges associated with sensorless control, aiming to reduce
Electric Scooters the reliance on sensors, which can be costly and vulnerable to faults. Sensorless control
techniques are investigated to estimate rotor position and speed accurately, enabling more
robust and fault-tolerant scooter operation.
Additionally, the paper discusses the impact of this fault-tolerant system on the overall
efficiency and reliability of two-wheeler electric scooters, making them a practical and
dependable choice for urban transportation. The findings presented in this research contribute
to the advancement of electric scooter technology, promoting sustainable and resilient
mobility solutions in today's environmentally conscious world.

Leveraging machine learning techniques, this project focuses on developing a system for
Automated
automated diagnosis of diseases such as cancer, pneumonia, or diabetic retinopathy by
Diagnosis of
127 analyzing medical images. The model aims to provide accurate and timely diagnoses,
Diseases using
assisting healthcare professionals in early detection and treatment planning.
Medical Imaging

Many cities and village areas have to install street lights. The amount of money spent is
enormous. Traditional sodium vapor lamps are very energy-consuming. The extra money can
come in handy for national development.

Solution: These IoT devices offer Automatic streetlight switching, lowering the cost of
IoT-Based Smart
maintenance, CO2 emission reduction, light pollution reduction, communication via wireless
128 Street Light
means, and power consumption. You, too, can build a prototype using only a few easily
System
accessible pieces of equipment.

This IoT Project Can Be Very Helpful For

Building a smart city

School of Computer Science and Engineering

66 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 School of Computer Science and Engineering
67 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 NLP for Kannada
language Large volumes of text data, including news articles, will have their sentiment analyzed using
129 sentiment analysis natural language processing (NLP) to reveal consumer and public opinion. More data sources
for news articles need to be created, trained, and utilized before this can be used for the Kannada language.
or social blogs

In geographic routing node needs to send their information to destination

node based on the geographic location here location information is used like address of the
node. In this way of geographic routing every node knows its own position and its nearby
node location. The frequent beaconing update reduce the network performance by increasing
traffic pattern it will increase the updates cost and decrease the routing performance. To
Enhanced position avoiding of this one in this paper we are projecting the (EAPU) Enhanced adaptive position
update using update with machine learning algorithm. EAPU follow two main principle (i) nodes moments
Machine leaning are frequent nature in adhoc network, it is hard to predict the moment and update their
130
approach in position frequently (ii) nodes which is moving towards the destination can update their
Mobile Adhoc position more frequently. By implementing machine learning algorithm various node
Network moments frequencies are analyzed. Based on the prediction ML algorithm the optimal path
can be chosen. By implementing these scenarios we need to use ns2 simulation using GPSR
(Greedy Perimeter Stateless Routing) protocol. This scenario is mathematically compared
with GPSR and periodic beaconing schemes shows that EAPU will increase the network
performance by reducing update cost, effective delivers of packets and average end to end
delay.

Machine learning (ML) has emerged as a transformative technology in healthcare, offering

Advancing
innovative solutions for disease diagnosis, treatment planning, personalized medicine, and
Healthcare
131 health management. This research aims to investigate the current landscape of ML
Through Machine
applications in healthcare, identify challenges, and propose future directions for research and
Learning:
implementation.

School of Computer Science and Engineering

68 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This research project introduces a groundbreaking approach to cybersecurity through the
integration of Reinforcement Learning (RL) with advanced network defense strategies.
Dubbed SentinelAI, our innovative framework empowers autonomous agents to adaptively
SentinelAI:
defend network infrastructures against emerging threats in dynamic environments. By
Reinforcement
leveraging RL's ability to learn optimal decision-making policies from experience, SentinelAI
Learning Agents
agents continuously evolve their defensive tactics to proactively identify, mitigate, and
132 for Autonomous
neutralize cyber attacks in real-time. Through a combination of deep neural networks, game
Cyber Defense in
theory, and network flow analysis, this study aims to develop a self-learning cyber defense
Dynamic
ecosystem capable of outmaneuvering even the most sophisticated adversaries. Our research
Networks
not only pushes the boundaries of AI-driven cybersecurity but also lays the foundation for a
new era of autonomous network defense, where human intervention is minimized, and
resilience is maximized against evolving cyber threats.

NLP stands for Natural Language Processing which is used to convert human language into
machine-understandable language. The inputs that we give to machines could be text, audio,
Natural Language
video, etc. For example, Siri which is used by Apple applications had a greater impact on the
processing for text
iPhone community. NLP is a combination of a large number of human visual systems, like the
classification by
133 phenomena including the human brain processing external information. AI is most important
using
for intelligent machines and Natural Language Processing is the fundamental tool. Intelligent
Convolutional
machines play a crucial role in performing various activities by reducing human efforts and
Neural Networks
error rates. Based on the research, this paper briefly discusses the concepts of deep learning
and NLP, various methods used for text generation models, etc.

Natural Language NLP stands for Natural Language Processing which is used to convert human language into
processing for text machine-understandable language. The inputs that we give to machines could be text, audio,
classification by video, etc. For example, Siri which is used by Apple applications had a greater impact on the
134
using iPhone community. NLP is a combination of a large number of human visual systems, like the
Convolutional phenomena including the human brain processing external information. AI is most important
Neural Networks for intelligent machines and Natural Language Processing is the fundamental tool.

SDNs decouple the control plane from the data plane, allowing centralized and programmable
network management, which presents unique opportunities for optimization. Performance
Performance
optimization of Software Defined Networks (SDNs) involves various strategies and
optimization of
135 techniques to enhance the efficiency, reliability, and scalability of network operations. AI and
Software Defined
machine learning algorithms can be utilized for predictive analysis, anomaly detection, traffic
Networks
prediction, and adaptive resource allocation, meeting the demands of modern network
environments.

School of Computer Science and Engineering

69 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Develop and LLMs have been released by various entities / research organizations for academic and
deploy a Large commercial use. These models can be used for generating human like responses on text based
Language Model systems with server-client model. The objective of the problem is to develop a tool for
(LLM) based tool network not connected to internet with following basic functions: (a) AI/ML based Text
for generating Summarization of given text. (b) Summarization of Science andamp; Technology
136 human like (Sandamp;T) related documents. (c) Summarization of NEWS papers headlines and editorial
responses to pages for quick overview of specific topics. (d) Reformatting and grammar checks with
natural language contextual integrity. (e) Additional capabilities which the developer can incorporate based on
inputs for network features available in open source LLM models. The solutions shall be graded on the
not connected capabilities, ease of use, flexibility and scalability of deployment and number of compatible
over internet models.

Cancer prediction using machine learning algorithms involves the analysis of medical data to
identify patterns indicative of cancerous conditions, aiming to enhance early detection and
improve patient outcomes. This process starts with collecting and preprocessing data from
various sources, such as medical records, genetic information, and imaging data. Feature
Cancer prediction
extraction and selection are crucial to highlight relevant indicators of cancer. Supervised
using machine
137 learning algorithms, including support vector machines, neural networks, and ensemble
learning
methods, are trained on labeled datasets to recognize cancerous patterns. These models are
algorithms
validated for accuracy and generalization before being deployed to predict cancer in new
patient data. Continuous refinement and integration of new data improve the model's
predictive capabilities, aiding healthcare professionals in making informed diagnostic
decisions and tailoring personalized treatment plans.

The security of RSA public key cryptosystem is based on the assumption that factoring of a
Hybrid large number (modulus) is difficult. In RSA if one can factor modulus into its prime numbers
Cryptography then the private key is also detected and hence the security of the cryptosystem is broken. The
Algorithm Based Rabin cryptosystem is an asymmetric cryptographic technique, whose security, like that of
138
on Modified RSA, is related to the difficulty of factorization. So in this topic a Rabin over RSA public key
Subset Sum cryptosystem called Hybrid cryptography Algorithm based on Modifieid subsetsum algorithm
Algorithm (HCAMSSA) is presented. This topic also presents comparison among Rabin, RSA and
HCAMSSA cryptosystems in respect of security and performance.

School of Computer Science and Engineering

70 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Investment is a business activity in which most people are interested in this globalization era.
There are several objects that are often used for investment, for example, gold, stocks and
property. In particular, property investment has increased significantly. Housing price trends
are not only the concern of buyers and sellers, but it also indicates the current economic
situation. There are many factors which have an impact on house prices, such as the number
Housing Price
of bedrooms and bathrooms. Even the nearby location, a location with great accessibility to
Prediction System
highways, expressways, schools, shopping malls and local employment opportunities
139 Using Linear
contribute to the rise in house price.
Manual house prediction becomes difficult, hence there are many systems developed for
Regression
house price prediction. We have proposed an advanced house prediction system using linear
regression. This system’s aim is to make a model which can give us a good house pricing
prediction based on other variables. We are going to use Linear Regression for this dataset
and hence it gives good accuracy.

Telegram Online Chatbots have been used in many fields ranging from education to
healthcare and are also used in e-commerce settings. This research aims at developing a web-
Deploying a
based chatbot for the Covenant Online General Store. The chatbot is developed using Python
Telegram E-
140 and React.js as the programming languages and MySQL (Structured Query Language) server
commerce website
as the database to give a structure to the e-commerce datasets and Admin Portal process. The
in cloud
e-commerce chatbot application for Covenant Online General Store seeks to provide an easy,
smart, and comfortable shopping experience for the Online Community.

Urban landscape and the physical challenges restrict the erection of small-scale wind
turbines. Large-scale turbines fared well compared to small-scale wind turbine models. Other
general challenges are as under: (1) Highly controlled energy sector (2) lack of awareness and
Development of
information (3) restricted access to technology (4) lack of competition (5) high transaction
141 Small Scale Wind
costs (6) poor market infrastructure (7) High investment requirements Expected Output: Eco-
energy device
friendly energy alternative, Unlike energy produced by fossil fuels while wind energy is
completely clean and eco-friendly. Not only is wind power free to harness, but its also 100%
renewable.

School of Computer Science and Engineering

71 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In the quest for improved patient outcomes, healthcare systems have predominantly relied on
big data, which consists of quantitative metrics and large-scale data analysis. However, these
approaches often fail to capture the nuanced, qualitative aspects of patient experiences that
Enhancing Patient are critical for delivering personalized and effective care. This project, "Enhancing Patient
Outcomes with Outcomes with Thick Data Insights in Healthcare," aims to integrate thick data
142 Thick Data methodologies into healthcare analytics to provide a comprehensive understanding of patient
Insights in behaviour’s, needs, and contexts, ultimately improving patient outcomes. Thick data
Healthcare encompasses rich qualitative insights derived from in-depth interviews, patient narratives,
ethnographic studies, and observational research. This project proposes a systematic
approach to incorporating thick data into healthcare decision-making processes alongside
traditional quantitative data.

Distributed denial-of-service (DDoS) attacks continue to trouble network operators and

service providers, and with increasing intensity. Effective response to DDoS can be slow
(because of manual diagnosis and interaction) and potentially self-defeating (as
indiscriminate filtering accomplishes a likely goal of the attacker), and this is the result of the
discrepancy between the service provider's flow-based, application-level view of traffic and
the network operator's packet-based, network-level view and limited functionality.
Furthermore, a network required to take action may be in an autonomous system (AS) several
Fixing DOS attack AShops away from the service, so it has no direct relationship with the service on whose
143 and DDOS attack behalf it acts. This paper presents Antidose, a means of interaction between a vulnerable
in IGMP peripheral service and an indirectly related AS that allows the AS to confidently deploy local
filtering with discrimination under the control of the remote service. We implement the core
filtering mechanism of antidose, and provide an analysis of it to demonstrate that conscious
attacks against the mechanism will not expose the AS to additional attacks. We present a
performance evaluation to show that the mechanism is operationally feasible in the emerging
trend of operators' willingness to increase the programmability of their hardware with SDN
technologies such as OpenFlow, as well as to act to mitigate attacks on downstream
customers.

School of Computer Science and Engineering

72 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The rising popularity of computing applications has piqued students' interest in computer
programming languages. Small errors are common when writing short code blocks due to a
Performance lack of knowledge and attention. Instead of merely providing syntax warnings, it would be
analysis of more beneficial to offer developers an Integrated Development Environment (IDE) that can
Retrieval- automatically correct short code blocks containing mistakes. This approach simplifies code
Augmented composition and makes it more time-efficient, thus enhancing the efficiency of large-scale
Generative (RAG) development environments. Given the increasing popularity of Python, our study aims to
144
model in improve the efficiency of writing Python code by providing an automatic code-correcting
automatic method. Additionally, automatic program evaluation has been implemented to assist in
program debugging small code blocks, ultimately aiding in the development of real-time computer
correction and applications. Recent advancements in Large Language Models (LLMs) have gained
evaluation significant traction in various NLP tasks, and the Retrieval-Augmented Generative (RAG)
model has shown great success in NLP applications. In this project, we will analyze the
performance of the RAG model for automatic program correction and evaluation.

School of Computer Science and Engineering

73 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In the modern era of digital health technology, the integration of artificial intelligence (AI)
with mobile applications has revolutionized healthcare delivery, making it more accessible,
efficient, and personalized. This project aims to develop a multilingual mobile application
Multilingual equipped with an AI-driven virtual health assistant to enhance healthcare accessibility and
Mobile quality for diverse populations worldwide.
145 Application for The proposed mobile application offers a user-friendly interface and supports multiple
AI-Driven Virtual languages, catering to the linguistic diversity of global users. The AI-driven virtual health
Health Assistant assistant serves as a personalized healthcare companion, providing users with real-time health
information, personalized recommendations, and assistance in managing their health and
wellness goals.

The integrity and security of voting systems are critical to ensuring fair and transparent
elections. Traditional voting systems are often vulnerable to fraud, manipulation, and cyber-
attacks. This project proposes the development of a blockchain-based secure voting system
designed to enhance the transparency, security, and trustworthiness of the electoral process.
By leveraging the decentralized and immutable nature of blockchain technology, the
Blockchain-Based proposed system ensures that votes are recorded accurately and cannot be tampered with.
146 Secure Voting Each vote is encrypted and added to the blockchain, creating a verifiable and auditable trail
System that can be independently checked by any party. The project will include a user-friendly
interface for voters, secure authentication mechanisms, and a robust backend infrastructure to
handle large-scale voting. The system will be tested in a simulated environment to
demonstrate its effectiveness and reliability. This innovative approach aims to address the
challenges associated with traditional voting systems and pave the way for more secure and
transparent elections

Software
Requirement
For any software development activities, there will be SRS document serves as source
Specification
document. This SRS document contains the requirements of a software to be developed. The
Document
business analyst will be discussing with the customer and understand the requirements to
Analysis and
prepare the SRS document. Once the SRS document is finalized, then it will be transferred to
147 Identification of
development team. If the developmental activities are taking place through cloud, the
important
development team has to identify the important parameters, resources and services required
parameters for
for the developmental activities. Automating the identification of important parameters
development
through machine learning algorithms is a challenging task.
through Machine
Learning

School of Computer Science and Engineering

74 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The rapid growth of the sharing economy has transformed the way individuals find and book
accommodations. Among the pioneering platforms, WanderLust has gained considerable
popularity as a trusted marketplace for short-term rentals. This project aims to enhance the
WanderLust experience by leveraging data-driven techniques to provide personalized
accommodation recommendations to users. Through the analysis of diverse data sources,
including user preferences, property attributes, location features, and guest reviews, our
An Artificial
system employs machine learning algorithms to generate tailored recommendations. By
Voice Assistant
considering various factors such as past user interactions, travel purpose, budget constraints,
designing an
148 and desired amenities, the proposed solution enables users to discover accommodations that
Endorse Web
best suit their needs and preferences. The main focus is on the company's operational
System
practices and its impacts on various domains. Four key topics are trust dynamics, house rules,
WanderLust
pricing strategies, and value creation. The most researched impacts are on urban tourism,
rental housing markets, tourist destinations, and hotels. Research on WanderLust's
operational aspects has grown significantly, with topics like trust, pricing, and impacts on
tourist destinations. Non-hospitality and tourism journals offer a broader coverage of
WanderLust-related topics.

School of Computer Science and Engineering

75 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Glaucoma is an eye disease that causes high pressure inside the eye. It is mostly caused by
damage to the nerve cells in the retina. If this problem isn't treated, it's very likely that the
person will go blind. Diagnosing glaucoma is currently done through procedures that take a
long time and must be carefully followed by following the enclosed directions. These tests
are very important for finding out when glaucoma starts and gets worse, but they are
complicated and take a lot of time, which can make eye care less accessible and less effective.
Hence current research develops the automated identification of glaucoma using deep
Automated
learning algorithms. The framework for automatically detecting glaucoma is made up of two
Glaucoma
critical parts. The first step is to use a Generative Adversarial Network (GAN). The main
Identification: A
purpose of this network is to create more image data that will improve the collection that will
GAN-EVCNN
149 be used for later classification processes. Utilizing GAN in the initial stage helps to increase
Framework for
the amount of image information that is available, which makes the base for the later stages
Accurate and
of the method of classification stronger and more complete. Then Enhanced version of
Efficient
convolutional Neural Network (EVCNN) is used for effective classification of glaucomatous
Diagnosis
images into affected and not affected. This architecture is specifically built to adeptly learn
complicated features from the input images, enabling it to recognize hidden trends and
ambiguities within the visual data. Utilizing an Enhanced version of convolutional Neural
Network (EVCNN) guarantees a comprehensive and subtle examination, resulting in the
production of extremely dependable results in the classification process. Also proposed
framework achieved the highest accuracy of 98.37% which considerably better outcome than
existing benchmark approaches.

School of Computer Science and Engineering

76 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Nowadays, next-generation networks such as the Internet of Things (IoT) and 6G are played a
vital role in providing an intelligent environment. The development of technologies helps to
create smart city applications like the healthcare system, smart industry, and smart water
BDN-GWMNN:
plan, etc. Any user accesses the developed applications; at the time, security, privacy, and
internet of things
confidentiality arechallenging to manage. So, this paper introduces the blockchain-defined
150 (IoT) enabled
networks with a grey wolf optimized modular neural network approach for managing the
secure smart city
smart environment security. During this process, construction, translation, and application
applications
layers are created, in which user authenticated based blocks are designed to handle the
security and privacy property. Then the optimized neural network is applied to maintain the
latency and computational resource utilization in IoT enabled smart applications

This project focuses on developing a sophisticated session-based recommendation system

utilizing natural language processing (NLP) to deliver real-time personalized suggestions.
Session-Based
Unlike traditional recommendation systems that rely on historical data, this approach
Recommendation
dynamically analyzes users' current interactions within a session to understand their
System
151 immediate preferences and intents. By employing advanced NLP techniques to process
Leveraging NLP
textual input and contextual cues, the system generates highly relevant recommendations on-
for Real-Time
the-fly. This innovation aims to significantly improve user engagement and satisfaction by
Personalization
providing timely and context-aware suggestions, thereby enhancing the overall user
experience across various digital platforms.

School of Computer Science and Engineering

77 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Cervical cancer is a significant cause of mortality among women worldwide, with early
detection being crucial for effective treatment and improved survival rates. Traditional
diagnostic methods, while effective, are limited by the need for expert interpretation, making
them time-consuming and susceptible to human error. Leveraging advancements in machine
learning (ML) and medical imaging, this research aims to develop an advanced ML
framework for enhancing cervical cancer detection from medical images. The proposed
framework addresses key challenges, including image quality and variability, feature
extraction and selection, data scarcity and annotation, and model interpretability and trust.

Our approach includes developing sophisticated image preprocessing techniques to

Enhancing standardize and enhance the quality of medical images and implementing advanced feature
Cervical Cancer engineering methods using Vision Transformers and transfer learning to accurately identify
Detection Using cancerous and precancerous conditions. To overcome the limitations of annotated data, we
152
Machine Learning explore data augmentation and semi-supervised learning strategies to expand the training
on Medical dataset effectively. Furthermore, the framework incorporates interpretability mechanisms to
Imaging ensure the ML models provide transparent and trustworthy results, facilitating their
acceptance in clinical settings.

The expected outcome is a robust ML system that improves the accuracy, efficiency, and
reliability of cervical cancer detection, ultimately contributing to better patient outcomes by
enabling early diagnosis and timely treatment. The system's performance will be evaluated
based on detection accuracy, image quality standardization, feature relevance, data utilization
efficiency, and interpretability, ensuring a comprehensive assessment of its clinical utility.
This research aims to revolutionize cervical cancer screening programs, reducing the
diagnostic burden on healthcare professionals and enhancing patient care through advanced
AI-driven solutions.

Transformer The study involves design and development of Deep Learning model for transforming the
Attention Model attention based gated inputs to gated outputs. This model involves various gates at different
153 for Identification levels and the same shall be used for representation of Machine understanding model. These
of a Person from Machine understanding model includes architecture that drives users from input images or
GAIT videos received for detection of classification of objects into different classes of objects

School of Computer Science and Engineering

78 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Extreme Neural Extreme Neural Network (xNN), a machine learning method for forecasting farmland
Network (xNN) emissions of greenhouse gases, was investigated in this research. Along with Carbon dioxide
based Predictive and Nitrous oxide fluxes, different ecological, agronomic, and environmental data were
154 Analysis of Green recorded at the investigation region throughout the course of a lengthy farming period. Cross-
House Emission validation and Statistical evaluation for estimating Carbon dioxide and Nitrous oxide fluxes
in Agricultural were used in the thorough study, which showed that the xNN model outperformed the
Field traditional ML models.

An effective ship torpedo defense system relies on 3D data about incoming torpedoes. To
fulfill this requirement, a Torpedo Detection and Location (TDL) sonar system was created.
This system features a transmitting transducer, two receiving arrays, and advanced signal
Under water
155 processing with 3D tracking capabilities. To evaluate the system's effectiveness, a test setup
Target Detection
for detecting and tracking torpedo-like objects in the air has been established. The hardware
used in this system is comparable to that of an acoustic tracking system used for surveying
pipelines on the seabed.
Smart EV
156 Smart EV Highways.
highways
157 NA NA

Reverse engineering plays a crucial role in various industries by enabling the analysis of
Securing Reverse
software, hardware, and systems to enhance understanding, improve security, and foster
Engineering
innovation. However, this practice often raises significant concerns about intellectual
Practices through
property (IP) theft and unauthorized replication. This project explores the integration of
Blockchain
blockchain technology to address these challenges, aiming to create a secure, transparent, and
Technology: A
158 tamper-proof system for managing reverse engineering activities.this project aims to provide
Decentralized
a robust solution for safeguarding intellectual property in reverse engineering, fostering a
Approach to
more secure and trustworthy environment for innovation and collaboration across industries.
Intellectual
By combining the strengths of reverse engineering and blockchain technology, the proposed
Property
framework promises to revolutionize how sensitive information is protected and managed in
Protection
the digital age.

An online answer correction system based on keywords would be an application that can
Online answer automatically identify and correct mistakes in answers or responses provided by users. The
159
script evaluation core functionality relies on natural language processing techniques and predefined sets of
keywords mapped to expected correct answers.

School of Computer Science and Engineering

79 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 As organizations increasingly adopt cloud computing to leverage its scalability, flexibility,
and cost-effectiveness, securing cloud-based networks has become a critical concern. Data
encryption emerges as a pivotal strategy to protect sensitive information from unauthorized
access, breaches, and cyber threats in cloud environments. This project explores
comprehensive data encryption strategies tailored to secure cloud-based networks, focusing
on both data at rest and data in transit.

The primary aim of this study is to develop and evaluate encryption methodologies that
ensure robust data protection while maintaining performance and usability. Key components
of the research include symmetric and asymmetric encryption algorithms, encryption key
management practices, and the integration of encryption with cloud services.
Data Encryption
For data at rest, the project examines the implementation of advanced encryption standards
Strategies for
160 (AES) and other symmetric encryption techniques to protect stored data. Emphasis is placed
Securing Cloud-
on seamless encryption and decryption processes to ensure data integrity and confidentiality
Based Networks
without significantly impacting system performance.

For data in transit, the study investigates the use of Transport Layer Security (TLS) and
Secure Sockets Layer (SSL) protocols to secure data transmitted between clients and cloud
servers. This includes analyzing the effectiveness of various cryptographic algorithms and
protocols in preventing eavesdropping, man-in-the-middle attacks, and data tampering.

A critical aspect of the project is the management of encryption keys, which are fundamental
to the security of encrypted data. The study explores key management solutions, such as
hardware security modules (HSMs), key management services (KMS) provided by cloud
vendors, and best practices for key generation, distribution, rotation, and revocation.

AI and IoT based

A lumpy disease is spreading very rapidly among the cattle, in this project we are developing
algorithm for
161 a novel technique to predict, monitor and alert the lumpy skin disease using deep learning
cattle lumpy
algorithms and IoT platform .
disease detection

Comprehensive In the digital age, social media platforms like Twitter have become critical channels for
Security communication, information dissemination, and social interaction.This project proposes the
162
Framework for development of a Comprehensive Security Framework for Twitter Data (CSFTD), designed
Twitter Data to enhance the security and integrity of user data on Twitter.

School of Computer Science and Engineering

80 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Most of the problems in the real world are multivariate i.e., involves many variables.
Multivariate data comprises of several datasets with more than one variable. Multivariate
datasets has power to change the use of data dramatically as database size increases and it
shows adequate results on predicting the effect on change in one variable will have on other
variable. These datasets consist of transitive and intrinsic hidden relationships among the
variables such as analyzing a variable is influenced by other process variables and
Multivariate Data
preferences. It is the situation where efficient multivariate data analysis techniques
Clustering using
exhaustively needed to catalog the given type of data. In the literature several techniques are
Heuristic Data
proposed and analyzed; one such technique is multivariate data clustering.
Intensive
This paper will present a unified framework of multivariate data clustering using heuristic
163 Computing and
data intensive computing and relevance feedback learning. The implementation starts with
Relevance
formalizing a heuristic data intensive computing (HDIC) which have the ability to handle data
Feedback
flows. Clustering the data is performed with proposed Relevance feedback learning algorithm
Learning
with consensus functions. These functions are selected as the change in the cluster ensemble
Approach
selection, combine and reduction. In this proposed approach we have used a new kind of
distance functions such as Camberra, Chi-square and Cityblock. The empirical analysis
shows that, the proposed approach attains better cluster ensembles on various multivariate
datasets taken from UCI and out performs with k-nearest neighbour (KNN) in different
settings. The performance of the proposed approach is assessed with Accuracy and F1-
measure.

Deepfake models may not effectively simulate natural blinking or eye movement. Analysing
these aspects can help identify manipulated videos. Use lip synchronization analysis to detect
discrepancies between audio and lip movements. Deepfake videos might have challenges in
accurately syncing generated faces with the audio track.Examine the video metadata for signs
of manipulation, such as inconsistencies in timestamps, editing software used, or other
anomalies . Analyse individual frames for artifacts, anomalies, or inconsistencies that may
indicate manipulation. Deepfake videos might exhibit unnatural artifacts, especially in high-
Deep Fake Video
motion sequences.Deepfake videos may exhibit variations in image quality, lighting, or
Detection using
164 resolution. Look for unnatural transitions between frames or inconsistencies in these aspects.
Deep Neural
Train deep learning models specifically for deepfake detection. Use pre-trained models or
Networks
create a custom model to classify videos as real or manipulated based on learned
features.Implement neural network architectures designed for deepfake detection, such as
using pre-trained models like OpenFace, DeepFace, or other deep neural networks.Train
models on diverse datasets that include both real and deepfake videos. This helps the model
generalize better to different manipulation techniques.Analyse the overall behavior and
movement patterns of individuals in the video. Deepfake models may struggle to accurately
simulate natural human behavior.

School of Computer Science and Engineering

81 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Real-Time Road The overarching goal of this research is to design, implement, and operationalize an
Accident innovative deep learning-powered system dedicated to the real-time detection of road
Detection Using accidents within CCTV footage. By leveraging cutting-edge technologies in deep learning,
165
Advanced Deep the intention is to markedly diminish false positive occurrences while simultaneously
Learning augmenting processing velocity. This dual enhancement is poised to revolutionize public
Architecture safety measures and elevate the efficiency of accident response mechanisms.

Generative AI image captioning represents a transformative technology for assisting visually

impaired individuals by providing textual descriptions of visual content. This study explores
the development and deployment of advanced generative AI models for image captioning,
leveraging deep learning architectures such as Convolutional Neural Networks (CNNs) for
image feature extraction and Recurrent Neural Networks (RNNs) or Transformers for
sequence generation. Our approach focuses on creating detailed, contextually accurate, and
user-friendly captions that enhance the accessibility of digital and real-world visual
information. By training the models on large-scale datasets containing diverse images and
corresponding captions, we aim to achieve high-quality, nuanced descriptions that convey
essential details, context, and emotions depicted in the images. The performance of our
models is evaluated using metrics like BLEU, METEOR, ROUGE, and CIDEr, as well as
Generative AI through user studies involving visually impaired participants to assess practical effectiveness
Image Captioning and user satisfaction. Preliminary results demonstrate that generative AI can produce highly
166
for Visually informative and contextually relevant captions, significantly aiding visually impaired users in
impaired people understanding and interacting with their environment. This research highlights the potential
of AI-driven image captioning as a crucial tool for enhancing digital accessibility and
promoting inclusivity. Future work will involve refining model accuracy, expanding dataset
diversity, and integrating real-time captioning capabilities to further support the needs of
visually impaired individuals.

School of Computer Science and Engineering

82 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The project aims to develop a precise and efficient system for segmenting blood cancer cells
in microscopic images through the application of fuzzy clustering methods. Accurate
Blood Cancer Cell segmentation of blood cells is crucial for diagnosing and treating blood cancers, such as
Segmentation leukemia, yet manual segmentation is labor-intensive and prone to subjectivity.
with Fuzzy This project employs fuzzy clustering algorithms, such as Fuzzy C-Means (FCM) , to
167
Clustering using partition microscopic images into distinct clusters representing different cell types. By
microscopic incorporating fuzzy logic, these techniques allow for the assignment of pixels to multiple
images clusters with varying degrees of membership, thereby capturing the inherent uncertainty and
ambiguity in image data.The project's outcome is a robust segmentation tool that facilitates
the accurate identification and quantification of blood cancer cells in microscopic images.

School of Computer Science and Engineering

83 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This study introduces a comprehensive approach to health monitoring, integrating
Comprehensive
Health psychological and physiological well-being through a multimodal system powered by
Monitoring: machine learning. By combining data from various sources such as wearable devices, mobile
Utilizing Machine applications, and psychological assessments, the system offers a holistic view of an
Learning for individual's health status. Leveraging advanced machine learning algorithms, including
168 decision trees, neural networks, and ensemble methods, predictive models are developed to
Psychological and
Physiological analyze and interpret the collected data. These models enable the identification of patterns
Well-being in a and correlations between psychological and physiological indicators, facilitating personalized
Multimodal interventions and recommendations for promoting well-being. The proposed system
System demonstrates promising results in enhancing health monitoring capabilities, emphasizing the
potential of machine learning in advancing holistic health management approaches.

Lung cancer is known as a malignant tumor that originates in the cells of the lungs, is often
associated with smoking, and is characterized by uncontrolled cell growth. Early detection of
LUNGCANARY-
PIONEERING lung cancer using clinical procedures is a very difficult undertaking that requires intrusive
EARLY LUNG techniques. Data on lung cancer might be gathered, and computational methods could be
CANCER utilized, improving lung cancer prediction and enabling early detection and prevention. In this
169 DETECTION study, we present the LungCanary approach, a pioneering method for addressing the global
USING health challenge of lung cancer through advanced machine-learning strategies. Our
MACHINE methodology involves a novel approach that effectively learns intricate patterns associated
LEARNING with early-stage lung cancer. Utilizing transfer learning principles, our novel model extracts
ALGORITHMS meaningful features from medical imaging data. To enhance diagnostic precision, Multiple
decision-making components are employed, each capturing different aspects of the data
patterns.

School of Computer Science and Engineering

84 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The project titled "System to Detect Fake Logos Online" aims to develop a sophisticated
detection system that utilizes image recognition and machine learning technologies to identify
counterfeit logos across various online platforms. This system is designed to combat the
widespread issue of brand impersonation and unauthorized use of logos, which can
significantly damage brand reputation and mislead consumers.

The core of the project involves training a deep learning model on a large dataset of authentic
and counterfeit logos, using convolutional neural networks (CNNs) to learn and differentiate
subtle variations in design that may indicate a fake. The system will integrate advanced
System to detect
170 algorithms for image analysis, ensuring high accuracy and the ability to operate in real-time
fake logos online
across diverse digital environments, including social media, e-commerce sites, and digital
advertisements.

This abstract would highlight the methodological approach, including data collection, model
training, and testing phases. It would also discuss the system's performance metrics, such as
accuracy, precision, and recall, demonstrating its effectiveness in detecting counterfeit logos.
Finally, the potential applications of this technology for brand protection agencies, legal
authorities, and businesses seeking to safeguard their intellectual property online would be
outlined, emphasizing its contribution to more secure digital commerce and brand integrity.

171 Green IoT Energy conservation in core IoT networks using reinforcement learning of the system.

School of Computer Science and Engineering

85 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 With the increasing importance of energy efficiency and sustainability, predicting and
optimizing energy consumption has become a critical area of research. This project aims to
apply data science and machine learning techniques to predict future energy consumption and
identify optimization strategies for reducing energy usage.

The project will begin by collecting a dataset containing historical energy consumption data
along with relevant features such as weather conditions, time of day, and occupancy levels.
The data will undergo preprocessing steps including cleaning, normalization, and feature
engineering. Various regression models, such as linear regression, decision trees, random
forests, and gradient boosting, will be employed to predict future energy consumption.
Energy
Consumption
To enhance prediction accuracy, the project will also explore time series forecasting methods
Prediction and
172 such as ARIMA, Prophet, and LSTM neural networks. Additionally, optimization algorithms
Optimization
will be implemented to suggest actionable strategies for reducing energy consumption while
Using Machine
maintaining operational efficiency.
Learning
The performance of the predictive models will be evaluated using metrics like Mean Absolute
Error (MAE), Mean Squared Error (MSE), and R-squared. Visualization techniques will be
used to present the predictions and optimization recommendations in an easily interpretable
format.

By the end of the project, students will gain hands-on experience in data preprocessing,
regression analysis, time series forecasting, and optimization techniques. They will also
develop an understanding of how data science can contribute to energy efficiency and
sustainability efforts in real-world applications.

School of Computer Science and Engineering

86 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In the manufacturing industry, equipment downtime can lead to significant financial losses
and operational inefficiencies. Predictive maintenance, powered by IoT and machine
learning, offers a solution to minimize unplanned downtime by predicting equipment failures
before they occur. This project aims to develop a predictive maintenance system for industrial
equipment, leveraging IoT sensors and machine learning algorithms.

The project will begin by installing IoT sensors on various pieces of industrial equipment to
collect real-time data on operational parameters such as temperature, vibration, pressure, and
usage hours. This data will be continuously transmitted to a central server or cloud platform
for storage and analysis.
Predictive
Machine learning models, including classification algorithms and anomaly detection
Maintenance for
techniques, will be employed to analyze the collected data. These models will be trained to
Industrial
173 identify patterns and indicators of potential equipment failures. Techniques such as Random
Equipment Using
Forest, Support Vector Machines, and neural networks will be explored to develop robust and
IoT and Machine
accurate predictive models.
Learning
The predictive maintenance system will also include a dashboard for real-time monitoring of
equipment health, alerting maintenance personnel to potential issues before they lead to
equipment failure. Recommendations for maintenance actions will be provided based on the
predictions generated by the machine learning models.

By the end of this project, students will gain practical experience in integrating IoT devices
with industrial systems, data collection and preprocessing, machine learning model
development, and implementing predictive maintenance strategies. They will also learn how
data science and IoT can be applied to enhance operational efficiency and reduce costs in
industrial settings.

School of Computer Science and Engineering

87 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This project focuses on optimizing urban traffic flow through the application of data analytics
and machine learning techniques. The project will involve collecting and analyzing traffic
data from various sources, including sensors, cameras, and GPS devices. By identifying
traffic patterns and congestion hotspots, the project aims to develop algorithms that can
predict and alleviate traffic congestion. The project will not only enhance the students' data
science skills but also contribute to reducing carbon emissions and improving the quality of
urban life. Traffic congestion is a major issue in urban areas, leading to increased travel
times, fuel consumption, and pollution. This project aims to tackle this problem by leveraging
data analytics and machine learning techniques to optimize urban traffic flow. The project
Urban Traffic
will involve collecting data from various sources, including traffic sensors, cameras, and GPS
Optimization
devices installed in vehicles.
174 Using Data
Analytics and
The collected data will be analyzed to identify traffic patterns, congestion hotspots, and other
Machine Learning
factors influencing traffic flow. Machine learning algorithms will then be applied to develop
models that can predict traffic congestion and suggest optimal routes for vehicles. These
models will take into account factors such as time of day, weather conditions, and special
events that may affect traffic flow.

By implementing these models in real-time traffic management systems, the project aims to
reduce congestion, shorten travel times, and improve overall traffic efficiency. Additionally,
the project will contribute to reducing carbon emissions and improving the quality of urban
life by promoting sustainable transportation practices.

School of Computer Science and Engineering

88 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 For neonatal development, there is a milestone that can last for up to two years. Babies can
reach full eyesight, speech, and physical development in two years. Milestone delays occur in
babies with Down syndrome, seizures within a year of life, and brain damage at birth.
Damage to the developing brain, which can occur during pregnancy or soon after delivery, is
the cause of cerebral palsy disease. Children are impacted in several areas, such as posture,
balance, reflex, muscle tone, coordination, control over their body movement, and reflex. To
DEVELOPING
meet developmental milestones, children with cerebral palsy require more than just normal
AN
speech and vision development. Infants gain from speech and vision stimulation as they
AUTOMATED
develop these abilities. Objects with lights, attention, novelty, focus, and movement are
VISION
essential for a baby's visual development. Colours are required by a baby between two and
STIMULATION
175 three months old. Do not shine lights directly into a baby's eyes. Amacrine and horizontal
TOOL FOR
cells support the Centre surround system, which helps to produce sharp images. Abnormal
CEREBRAL
ocular exposure has an impact on both horizontal and amacrine cell formation. The Visual
PALSY
System is used to identify possibilities, communicate with carers, improve baby
REHABILITATIO
concentration, and provide external visual experiences.
N
To encourage good speech development, babies should be taught how to focus on lip
movement, detect communicative behavior, and learn how to use, play with, and understand
the meaning of objects through sound toys. As a result, children with cerebral palsy require a
specialized training programme for speech and ocular stimulation, which is not currently
available on the market. Thus, the project activity intends to provide training tools for
children with cerebral palsy.

Energy This project aims to develop a machine learning-based system for forecasting energy
Consumption consumption in smart grids. By analyzing historical energy usage data, weather patterns, and
176
Forecasting for other relevant factors, the model will predict future energy demand, facilitating efficient
Smart Grids resource allocation, and grid management.

School of Computer Science and Engineering

89 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This project focuses on implementing a precision agriculture monitoring system using drones
and unmanned aerial vehicles (UAVs). The system will utilize drones equipped with various
sensors, such as multispectral cameras and thermal imaging sensors, to collect data on crop
health, soil moisture levels, and pest infestations.

The collected data will be analyzed using data analytics and machine learning algorithms to
provide farmers with actionable insights. These insights can include optimized irrigation
schedules, early detection of crop diseases, and targeted pest management strategies. By
leveraging drone technology and data-driven analytics, this project aims to enhance
agricultural productivity, reduce resource wastage, and promote sustainable farming practices.

Precision agriculture, enabled by advanced technologies like drones and unmanned aerial
Precision
vehicles (UAVs), has revolutionized farming practices by offering detailed insights into crop
Agriculture
health, soil conditions, and environmental factors. This project aims to develop a
Monitoring Using
177 comprehensive precision agriculture monitoring system that leverages drones and UAVs to
Drones and
collect high-resolution data for analysis.
UAVs: A Data-
Driven Approach
The drones will be equipped with a range of sensors, including multispectral cameras,
thermal imaging sensors, and GPS modules, to capture data such as crop health indicators
(NDVI), soil moisture levels, and temperature gradients. This data will be transmitted in real-
time to a central database for analysis.

Data analytics and machine learning algorithms will be applied to analyze the collected data
and generate actionable insights for farmers. These insights may include optimized irrigation
schedules based on soil moisture levels, early detection of crop diseases through analysis of
spectral signatures, and identification of pest infestations through thermal imaging.

By providing farmers with precise and timely information, this project aims to enhance
decision-making processes, increase crop yields, and reduce resource wastage. Furthermore,
The advances in machines and technologies used in smart farming using machine learning,
useful and accurate information about different matters plays a significant role in it. It focuses
Smart Farming
on predicting the appropriate crop based on the climatic situations and the yield of the crop
178 using Machine
based on the historic data by using supervised machine learning algorithms. The only remedy
Learning
to the crisis is to do all that is possible to make agriculture a profitable enterprise and attract
the farmers to continue the crop production activities.

School of Computer Science and Engineering

90 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The primary concern in preventing unauthorized behavior in computer systems is the
automatic detection of intrusions. Adopting intrusion detection systems on hosts and
networks necessitates having a wide understanding of computer security. Misuse detection
and signature-based detection are the two basic methods that underpin the majority of IDS
Attack detection and IPS systems. It establishes a range of "unacceptable" behaviors and sounds an alarm
179 by deep learning when the system exhibits any of them. Signature-based intrusion detection systems (IDS)
techniques make it simple to identify frequent attempts and offer defenses against them by matching
signatures or matching string patterns. But current signature-based detection is ineffective in
the current environment, when fresh intrusions/attempts are recorded nearly every day.
Numerous IDPS have been suggested, yet they are all deficient in some point. Here by using
machine learning and deep learning, threat can be detected and prevented.

To ensure that a design can be manufactured or integrated as smoothly as possible, there

are several considerations to keep in mind. Indoor agriculture lighting and sensors have
specific requirements that need expertise outside of commercial and industrial lighting
designed for humans. Vertical farming is a popular way of producing vegetables indoors
under tightly regulated weather and light conditions. The goal behind the concept is to
Digitally deploy automated facilities that, with the help of technology, strive to have the least
automated vertical possible impact on the environment while significantly increasing agricultural product
180 farming production. This form of agricultural operation has the distinct benefit of not requiring
Geoscience Earth large horizontal flat expanses. The word quotverticalquot is used since this manufacturing
Science may be done in floors that is, piled on many planes. Vertical farms use sustainable energy
sources, artificial light management, and careful use of water and other precious natural
resources throughout the planting process.
Environmentally regulated farming technology, such as temperature and humidity
management, may and should be used in vertical farms.

School of Computer Science and Engineering

91 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 This paper proposes a comprehensive solution to mitigate the persistent issue of crop damage
caused by wild animals, particularly elephants, in rural areas. The conflict between farmers
and wildlife poses significant challenges to agricultural livelihoods and conservation efforts,
necessitating innovative approaches to address human-wildlife conflict. Our solution, the
Smart Wildlife Deterrence System, integrates advanced technologies such as sensors,
"Smart Wildlife cameras, and artificial intelligence algorithms to detect and deter wild animals from entering
Deterrence agricultural fields. The system utilizes perimeter sensors to detect the presence of wildlife
System: and triggers automated responses to deter them effectively.
Mitigating Crop Furthermore, the Smart Wildlife Deterrence System employs non-lethal deterrents such as
181
Damage in Rural sound and light signals, as well as physical barriers activated by the detection system, to
Areas through create deterrent zones around crop fields. Real-time monitoring and data analytics capabilities
Technological enable proactive intervention and adjustment of deterrence strategies based on wildlife
Intervention" behaviour patterns. By deploying this technology-driven approach, rural communities can
reduce crop losses, mitigate human-wildlife conflict, and promote coexistence between
farmers and wildlife. The Smart Wildlife Deterrence System offers a scalable and sustainable
solution to safeguard agricultural livelihoods while conserving wildlife populations in rural
landscapes.

Wireless sensor networks, due to their nature, are more prone to security threats than other
networks. Developments in WSNs have led to the introduction of many protocols specially
Secure developed for security purposes. Most of these protocols are not efficient in terms of putting
Knowledge and an excessive computational and energy consumption burden on small nodes in WSNs. This
Cluster-Based paper proposes a knowledge-based context-aware approach for handling the intrusions
Intrusion generated by malicious nodes. The system operates on a knowledge base, located at the base
182
Detection station, which is used to store the events generated by the nodes inside the network. The
Mechanism for events are categorized and the cluster heads (CHs) are acknowledged to block maliciously
Smart Wireless repeated activities generated. The CHs can also get informational records about the
Sensor Networks maliciousness of intruder nodes by using their inference engines. The mechanism of events
logging and analysis by the base station greatly affects the performance of nodes in the
network by reducing the extra security-related load on them.

AI-Enhanced E-
Learning Platform Education is a fundamental right, but traditional e-learning platforms often fail to meet the
183 for Visually and needs of visually and cognitively impaired individuals. This project aims to develop an AI-
Cognitively enhanced e-learning platform designed specifically for users with visual and cognitive
Impaired Users impairments, ensuring an inclusive and effective learning experience.

School of Computer Science and Engineering

92 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 n traffic light design, modeling and verification are crucial steps because they verify the
accuracy of the model and ascertain whether or not it has desirable characteristics. Before
Modeling and being used, the traffic light design models need to be confirmed. The traffic light model needs
Verification of to be carefully verified, particularly in synchronized traffic light intersection systems where
Traffic Light the traffic lights at several crossings require numerous communication channels for the
184
Signal using synchronization of the message among intersections and traffic light groups. In this study, the
Uppal and Timed use of Uppaal time automata for the modeling and verification of traffic-right design models
Automata was proposed. Our models and templates assist modellers in developing traffic-right designs
in a formal model and can confirm that the model's safety features and structure are accurate.

IoT for The objective of this proposal is to enhance guest experiences and operational efficiency in
hospitality: the hospitality sector by incorporating IoT technologies. Using their cellphones, guests can
Improving adjust the hotel's settings with smart room controls. Asset monitoring systems monitor hotel
185
Operational assets with the use of sensors, which improves supply chain visibility and lowers theft. In
Efficiency and order to provide tailored recommendations and raise the standard of service, guest analytics
Guest Experience systems gather information on visitor behavior.

Facial The attendance system these days has become havoc and not done correctly. For this problem
Recognition we used CNN to locate where the student has come to the class or not is the main criteria. By
186
Attendance Based doing this we can solve the issue as well as save a lot of time for the faculty who will be
System having extra time to take. This can be used in various domains as well.

Gen AI based assessments are the future. Evaluation of such assessments are truly not figured
Formulation of and remains as a challenge. Figuring out the various techniques of assessments based on
Gen AI based formulated hypotheses using a collection of multimodal data can help generate assessment
187
assessment ideas that can be used in the future. This will require data collection, analysing the collected
techniques data, figuring out correlations, generation of hypotheses, validation of hypotheses, training
and testing using the obtained parameters.

Smart farming is a device built to detect the soil's temperature, moisture, and humidity. This
Smart Farming device will help the farmer to know the conditions of the soil. We need an Arduino Nano
188 Using IoT IDE, temperature and moisture sensors, and some jumper wires to build this device. This
Technologies device will have a direct content with the soil, and the measurement is shown on the monitor
screen.

School of Computer Science and Engineering

93 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 An artificial intelligence (AI) system aids autism treatment and early intervention. When
robots are introduced to people with autism, social and communication skills can grow. The
social, educational, and assistive robots are the robotics variations that help autism. The
Artificial
expected goal is to use emotional intelligence to control the emotions of autistic people
Intelligence for
towards others and to comprehend the emotions of others. Robots' ability to enhance
Facial Emotion
189 emotional intelligence in individuals with autism spectrum disorder (ASD). In this research
Recognition in
paper, the FER2013 dataset is used. The comparison study uses different deep-learning neural
Autism Support
network architectures to recognize facial emotions with high accuracy to build a generic best
Systems
model to deploy in real-world scenarios to learn about human emotions.

This application is helpful for both patient and doctor book and schedule appointments. The
doctors can enable empty slots available, and the patient can book their slots accordingly.
Patient and
Also, this system can help to store the medical history of the patient. So, whenever patients
190 Doctor Handling
login into the system, they can look into the prescription suggested by the doctors. Also, the
System
doctor can see those reports to understand the improvements in the patient body from
previous tests and reports. It is easily manageable, and patients don’t need to wait for doctor
appointments.

School of Computer Science and Engineering

94 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Distributed denial of service (DDoS) attacks is rising exponentially and creates a severe
threat to security. Generally, the DDoS attack may appear uncomplicated but they are hard to
prevent and considered as one of the most significant cybersecurity issues.
The main purpose of this work involves the optimal detection of data samples as normal data
samples and malicious/attacked data samples. This paper proposes a security algorithm
against DDoS attacks by employing four significant phases namely the Database training
phase, Data pre-processing phase, Feature selection phase and Classification phase.
CNN based
Initially, the data samples are to be trained before using it for attack detection. Later, the
classifier for
sample group is created for every file and the data samples are pre-processed in the data pre-
191 detecting DDoS
processing phase. Secondly, in feature selection phase, the selected features are optimized by
attack in cloud
employing kernel principal component analysis to obtain optimal features.
computing
In the classification phase, a Convolutional Neural Networks -based discrete elephant
herding optimization (CNN-DEHO) classifier is utilized to detect the data sample as normal
data and attacked or malicious data. Finally, the proposed approach is examined for two
different databases namely NSL-KDD, UNSW-NB15. The experimental analyses will be
performed for various simulation metrics and the outcome reveals that the detection system
performances are high using CNN-DEHO approach than other approaches.

The rapid growth of data in digital environments has heightened the need for efficient and
accurate recommendation systems. Graph Neural Networks (GNNs) have emerged as a
Multi-Level powerful tool for leveraging the rich relational structures inherent in user-item interactions.
Graph Neural This paper explores the application of hierarchical GNN-based approaches for
192 Networks for recommendation systems, aiming to enhance their accuracy and scalability. By organizing
Personalized users and items into hierarchical structures, our method captures multi-level relationships and
Recommendations dependencies that traditional GNNs might overlook. The hierarchical framework enables the
model to aggregate information at different levels of granularity, providing a more
comprehensive understanding of user preferences and item characteristics.

School of Computer Science and Engineering

95 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The Building Management System (BMS) project addresses the critical need for efficient and
sustainable management of commercial and residential buildings. By developing a
Building comprehensive digital platform that integrates monitoring, energy management, maintenance,
Management security, and environmental monitoring functionalities, the project aims to enhance
System (BMS): operational efficiency, reduce costs, and promote environmental sustainability within
193 Enhancing buildings. Through meticulous planning, systematic design, rigorous development, and
Operational comprehensive evaluation, the BMS project endeavors to deliver a robust and user-friendly
Efficiency and solution that meets the diverse needs of building owners, managers, and occupants. The
Sustainability successful implementation of BMS holds the potential to transform building operations,
optimize resource utilization, and create healthier and more comfortable indoor environments
for occupants while advancing sustainability goals.

Abstract:
Urban traffic congestion is a significant problem affecting millions of commuters daily. This
project focuses on developing a real-time traffic prediction and congestion management
Real-Time Traffic
system using big data analytics and machine learning. The system will collect and process
Prediction and
data from various sources, including traffic cameras, GPS devices, social media, and weather
Congestion
194 reports. Advanced machine learning models, such as neural networks and ensemble methods,
Management
will be used to predict traffic flow and identify potential congestion hotspots. The project will
Using Big Data
also explore the use of reinforcement learning to optimize traffic signal timings dynamically.
Analytics
The outcome will be a comprehensive traffic management system capable of providing real-
time traffic forecasts and actionable insights to reduce congestion and improve urban
mobility.

As everyone likes to enjoy watch TV during rainy time. But because of rain , the signals will
Rain Proof TV
195 lost by dish antenna and TV cant show any channels . To overcome this an innovative has to
streaming
be found and formulated in this project .

School of Computer Science and Engineering

96 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In the current digital landscape, cyber-attacks are evolving at an alarming pace. These attacks
target vulnerabilities in networks, aiming to steal sensitive data, disrupt critical operations, or
extort money. The ever-growing number of cyber-attacks poses a significant threat to network
security, potentially leading to data breaches, financial losses, and operational disruptions.
The consequences of successful cyber-attacks can be devastating, leading to financial losses,
reputational damage, and even threats to public safety. To combat this evolving threat
landscape, Intrusion Detection Systems (IDS) serve as a vital line of defense. IDS
continuously monitors network traffic for suspicious activities that deviate from normal
patterns. By identifying potential intrusions and alerting security personnel, IDS enable
timely intervention and mitigation strategies. However, the effectiveness of IDS hinges on
Boosting
their ability to accurately detect an ever-expanding range of attack techniques. Intrusion
Accuracy Values
Detection Systems (IDS) play a critical role in safeguarding networks by identifying and
Using Multiple
196 alerting security personnel to malicious activity. However, the effectiveness of IDS hinges on
classification
their ability to accurately detect an ever-expanding range of attack techniques. Machine
algorithms in
Learning algorithms offer promising solutions for building robust IDS models. This research
Kyoto Dataset
investigates the potential of different machine Learning algorithms to get the best accuracy.
Traditional benchmark datasets, like KDD Cup 99, have been criticized for not reflecting the
complexities of modern network traffic. This limitation can hinder the development and
evaluation of robust IDS models employing machine learning algorithms. To address this, we
leverage the Kyoto 2006+ dataset, a collection of labeled network traffic encompassing
normal, known attack, and unknown attack instances. The effectiveness was evaluated using a
comprehensive set of metrics, including True Positve Rate(TPR), True Negative Rate (TNR),
Precision, and False Discovery Rate (FDR). The Random Forest algorithm achieved the
highest accuracy of 98.75% with True Positive Rate of 99.98% and False Discovery Rate of
almost 0 (0.0000138).

Lung cancer death rates have been rising in recent years, making it imperative to determine if
a tumor has progressed to cancer or not. Many lives can be saved if an accurate diagnosis is
made early on, and it can also assist medical professionals in initiating treatment. By
determining the tumor's location, size, and other characteristics, computed tomography is
LUNG CANCER
essential in ensuring that tumor conditions are met. This early cancer detection helps save
PREDICTION
many lives that might otherwise be lost. The pre-processed image goes through a
197 USING
segmentation step before being sent on to feature extraction. The extracted features are then
MACHINE
trained using machine learning classification methods such as Random Forest, SVM (Support
LEARNING
Vector Machines), and ANN (Artificial Neural Network). A diagnosis of benign or malignant
tumor is determined based on the categorization findings received. To find out which
algorithm has the best prediction accuracy, several characteristics are measured, such
accuracy, recall, and precision.

Artificial intelligence (AI) technology holds both great promise to transform mental
Artificial
Intelligence for healthcare and potential pitfalls. This article provides an overview of AI and current
198 Mental Health and applications in healthcare, a review of recent original research on AI specific to mental
School of Computer Science and Engineering
97 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Mental Illnesses: health, and a discussion of how AI can supplement clinical practice while considering its
A Case Study current limitations, areas needing additional research, and ethical implications regarding AI
technology.

School of Computer Science and Engineering

98 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Climate change
199 Climate change impact prediction
impact prediction

It uses machine learning algorithms to correctly identify 99 plant species through the binary
leaf images and evaluated features. These features include shape, margin, and texture. We
Plant Species
find how the leaves are because their volume, prevalence, and unique characteristics can
200 Identification
serve as an effective measure to identify plant species. We use implementation of
using Machine
classification machine learning algorithms and also learn to benchmark the significance of
different classifiers in image classification problems.

This project proposes the development of an innovative Tablet Monitoring System (TMS)
leveraging Internet of Things (IoT) technology. The TMS aims to provide real-time
"IoT-based Tablet
monitoring and control functionalities for various applications, including healthcare,
Monitoring
education, and industrial processes. By integrating IoT sensors and communication protocols,
201 System for
the system enables remote monitoring of tablet devices, ensuring optimal performance,
Enhanced Control
security, and user experience. This project presents the design, implementation, and
and Efficiency"
evaluation of the IoT-based TMS, highlighting its potential benefits and applications across
different sectors.
CROP YIELD
Indian agriculture, a backbone of the country's economy, faces challenges such as climate
PREDICTION
variability, soil degradation, and resource constraints. Accurate crop yield prediction can
BASED ON
202 significantly enhance decision-making for farmers, policymakers, and stakeholders.
INDIAN
Leveraging Artificial Intelligence (AI) and Machine Learning (ML) for crop yield prediction
AGRICULTURE
provides precise and actionable insights, aiding in sustainable agricultural practices.
USING AIML

Abstract:

The rapid proliferation of Internet of Things (IoT) devices in various sectors has sparked a
surge in demand for efficient and reliable communication technologies. As the fifth-
generation (5G) cellular network emerges as a key enabler for IoT applications, the design
Design and and implementation of antennas tailored for such applications have become crucial. This
implementation of paper presents an overview of the design considerations, challenges, and recent advancements
203
Antennas for 5G in antennas for 5G IoT applications. It discusses key factors such as antenna size, bandwidth,
IOT Applications. efficiency, and radiation patterns, along with techniques to enhance antenna performance in
diverse IoT scenarios. Furthermore, it explores innovative antenna designs, such as compact,
multiband, and reconfigurable antennas, optimized to meet the stringent requirements of 5G
IoT deployments. Through a comprehensive analysis of existing literature and case studies,
this paper aims to provide insights into the evolving landscape of antennas for 5G IoT
applications and pave the way for future research endeavors in this domain.

School of Computer Science and Engineering

99 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 The problem of distributed estimation over wireless sensor networks in an adversarial
environment with the presence of attacks on sensed and communicated information is
considered. To tackle with this problem, a secure diffusion leastmean squares (S-dLMS)
algorithm is proposed. The proposed SdLMS can be considered as a hybrid system, which
consists of a non-cooperative LMS (nc-LMS) subsystem and a diffusion LMS (dLMS)
Secure Distributed subsystem. The nc-LMS subsystem is used to provide a reliable reference estimate, which is
Estimation over further used for constructing the threshold test to detect the trust neighbors of each node.
204 Wireless Sensor Then, based on the detected secure network topology, the dLMS subsystem is performed by
Networks under combining the received information from the trust neighbors. The performance of the
Attacks proposed SdLMS algorithm in the mean and mean-square senses is analyzed, and then an
adaptive rule is suggested to select the threshold for detection. Finally, some simulations are
performed to show the effectiveness of the proposed S-dLMS algorithm under fixed and time-
varying attacks, respectively.

The most important thing when it comes to cyber security is having a strong password. To
determine how strong your passwords are, you can create a cyber security project that follows
the OWASP (Open Web Application Security Project) guidelines. The algorithms used in
machine learning are used to train the application. The system is given datasets containing
Password Strength
205 weak passwords, passwords from previous breaches, etc. Train your system with a dataset
Tester
such that the current user enters a password that seems similar, and the system alerts them.
The program should establish fundamental guidelines such as a minimum password length,
complexity, and exclusion of username or other publicly accessible information, as well as
the usage of special characters and numbers.

The project aims to develop a mobile application that provides users with a comprehensive
overview of all installed applications on their device and empowers them to manage internet
access permissions on an individual basis. In an era where privacy and data security are
paramount concerns, this application serves as a crucial tool for users to exercise greater
control over their digital footprint.
206 APPWARDEN
The application will offer a user-friendly interface displaying a list of installed applications,
along with pertinent information such as data usage statistics and network access
permissions. Users will have the ability to selectively enable or disable internet access for
each application, thereby enhancing privacy and security while optimizing data usage.

School of Computer Science and Engineering

100 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 School of Computer Science and Engineering
101 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 In the digital age, sharing lengthy and complex URLs can be cumbersome and inconvenient,
particularly in contexts where brevity and clarity are essential, such as social media, emails,
and messaging applications. This project introduces the development of a "URL Shortener"
application, aimed at converting long URLs into short, easily shareable links while
maintaining the ability to redirect users to the original web addresses.

The URL Shortener application leverages a simple and efficient algorithm to generate unique,
shortened URLs that map to the original long URLs. Key features include a user-friendly
interface for URL input and shortening, a database for storing URL mappings, and robust
redirection functionality. Additionally, the application incorporates features for tracking the
207 URL Shortener number of clicks on shortened URLs and providing users with analytics on their usage.

The backend of the application is built using a modern web framework, ensuring scalability
and performance, while the frontend offers a clean and intuitive user experience. Security
measures, such as rate limiting and validation, are implemented to prevent abuse and ensure
the reliability of the service.

This project aims to enhance the convenience of sharing URLs by providing a reliable and
efficient URL shortening service. By doing so, it addresses the need for simplicity and
accessibility in URL management, making it easier for users to share links across various
platforms without the limitations imposed by long URLs.

School of Computer Science and Engineering

102 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 Evolutionary Computation is an optimization process to get the best solution, out of the
randomly found solution candidate space or pool. Evolutionary principles are based on the
survival of the fittest biological principles. In Evolutionary Computation we map the data in
the data mining to the genotype and phenotype genetic algorithm variants. The fitness of the
genetic code determines the fitness
of the child through the parents, which states that the fitness of the data mining relies on the
fitness of the data under consideration. In this regard we use genetic algorithms in the
consciousness based neural network known as neuro-evolution to determine the fitness of the
data at hand based on the prediction accuracy levels to determine the overall fitness of the data
under consideration for data mining. This is achievable through the concept of mutation of the
data. The mutations of the human DNA, which contain the genetic data of the individual
is a key stakeholder in the quality of the data, which means that mutations through variation
and selection lead to better individuals or better data, that is ultimately required for data
mining to be more successful and more accurate.

NEURO-
EVOLUTION: A
TREASURE FOR
208 DATA MINING
USING
EVOLUTIONAR
Y MUTATIONS.

The stimulus to carry out this research was to identify and propose a secure framework for
the Internet of Things (IoT). Due to the massive accessibility and interconnection of IoT
devices, systems are at risk of being exploited by hackers. Therefore, there is a need to find

School of Computer Science and Engineering

103 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 an advanced security framework that covers data security, data confidentiality, and data
Advanced
integrity issues. The study uses a systematic literature review (SLR) technique and complete
Security
substantive literature is reviewed to find out the constructs and themes in the existing
209 Framework for
literature. We performed it in four steps, which were inclusion, eligibility, screening, and
Internet of Things
identification. We reviewed around 568 articles from well-reputable journals, and after
(IoT)
exclusion, 260 articles and 54 reports were analyzed. We performed an analysis using
MAXQDA in which the nodes and themes were first identified. After the classification, a
qualitative model was generated using MAXQDA. The proposed model is supported by the
literature so it will be useful for the IT managers, developers, and the users of IoT.

School of Computer Science and Engineering

104 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in
 An IoT-Based Effective and real-time eyeblink detection is of wide range applications, such as deception
Intelligent System detection, drive fatigue detection, face anti spoofing. Despite previous efforts, most of
for Real-Time existing focus on addressing the eyeblink detection problem under constrained indoor
210
Parking conditions with relative consistent subject and environment setup. Nevertheless, towards
Monitoring and practical applications, eyeblink detection in the wild is highly preferred, and of greater
Automatic Billing challenges.

Recruiters from companies and HR need help reviewing many resumes whenever a job opens.
In cases of high demand for job roles, many job applications come flowing in. Sometimes,
when skimming through resumes, there is a possibility that an ideal candidate’s resume does
211 Resume Parser not receive the necessary attention, or maybe it is missed due to the enormous pile of
applications. That makes things difficult for the job applicants and the company where they
would have been more suited to work. It is a good application for machine learning, as it can
help people browse through resumes.

This project aims to develop an intrusion detection system (IDS) that leverages machine
learning to identify and prevent unauthorised access to networks. The system analyses
Intrusion network traffic data to detect anomalies that may indicate malicious activities. By training on
Detection System both normal and malicious traffic patterns, the model learns to differentiate between
212
Using Machine legitimate and suspicious behaviour. The IDS can then alert administrators or automatically
Learning block potential threats in real-time. This machine learning-based approach enhances the
accuracy and responsiveness of traditional intrusion detection systems, providing better
protection for network security.

Placement Coordinator Dr. Geetha G

School of CSE Director – SCSE

School of Computer Science and Engineering

105 of page 98
Faculty of Engineering & Technology, 45th km, NH - 209, Jakkasandra Post
Bangalore - Kanakapura Main Road, Ramanagara District - 562 112
www.set.jainuniversity.ac.in | www.jainuniversity.ac.in