Capstone projeCts 2024

Project Title
Group Number Project Description
Definition and Solution: Communication is a fundamental aspect of human interaction, yet individuals who use sign language face significant challenges in bridging the
gap with those who rely on spoken or written languages. Current solutions struggle with accurate gesture recognition and real-time translation. Our project aims to
build a real-time accurate system that would help communicating with the disabled by translating sign language into audio and text and vice-versa.
Outcome: A system that will accurately translate sign language to text/audio in realtime. For communicating with the disabled the system will enable us to define
gestures with respect to the input text/audio data.
SignLingo- Sign Language Translation Objectives: Accurate Gesture Recognition, Real-Time Translation, NLP Integration, Text to speech & vice-versa, Continuous Learning & Adaptation, Cross Model
CPG 1 and Communication System Integration.
Our Project will be a groundbreaking mobile application designed to detect dyslexia in children aged 3-5 through engaging quizzes and a cutting-edge machine learning
model. The app will boast a user-friendly interface with age-appropriate quizzes featuring multimedia elements, ensuring an interactive experience for young users. The
heart of the solution lies in a sophisticated machine learning algorithm . By analyzing responses to quizzes, the model can identify patterns indicative of potential
dyslexic traits, taking into account factors like response time, accuracy, and linguistic cues. Our App will go beyond identification, providing personalized feedback to
parents or guardians with insights into potential areas of concern and recommendations for supporting the child's language development. The app will also feature
progress tracking. With accessibility features catering to different learning styles, our app aims to empower parents, facilitate early intervention, and contribute to
CPG 2 Dyslexia Detection ongoing research in the field. Ultimately, it aspires to positively impact young children's lives by ensuring timely support for successful learning experiences.
The Capstone Project Submission Portal aims to revolutionize the traditional approach to managing and evaluating capstone projects of the computer science and
engineering department and beyond. By addressing key objectives such as streamlined registration, comprehensive evaluation processes, multimedia integration, and
Revamp Capstone Portal: One stop robust analytics, this portal will contribute to a more seamless and impactful capstone project experience for both students and faculty members. The inclusion of a
CPG 3 capstone solution dedicated mobile application further enhances accessibility, ensuring that users can engage with the platform anytime, anywhere.
CPG 4 Plastic detection To detect plastic in water bodies using ML,DL and IOT ideas

Introduction:
The Face Recognition Hostel Attendance System is a modern solution designed to enhance attendance tracking in hostels. By leveraging advanced face recognition
technology, the system aims to provide accurate, real-time attendance records, improve security, and streamline administrative processes.

Objectives:

Accurate Tracking: Implement a reliable face recognition system for precise attendance monitoring.
Real-time Access: Provide administrators with instant access to attendance data via a user-friendly interface.
Enhanced Security: Utilize face recognition to ensure secure access control and prevent unauthorized entry.
Streamlined Processes: Automate attendance tracking to reduce administrative workload and optimize efficiency.

Solution Proposed:

Face Recognition Module:

Utilize cutting-edge facial recognition algorithms for accurate resident identification.
Deploy high-resolution cameras at entry points for optimal data capture.
Database Management System:
Maintain a centralized database with secure storage for resident profiles and attendance records.

User Interface:
Develop an intuitive web-based interface for administrators.

Automation System:
Implement automated attendance tracking using face recognition technology.

Security Features:
Incorporate encryption and anti-spoofing measures to secure data and prevent unauthorized access.
Face Recognition Hostel Attendace
CPG 5 System Outcome:
Visual odometry (VO) is a technique used to estimate the motion of a camera or a vehicle by analyzing a sequence of images. It is widely applied in various fields due
Stereo Camera Based Visual to its ability to provide relative position and orientation information without relying on external systems like GPS. In this project, we will explore stereo camera based
CPG 6 Odometry visual odometry targeting applications in Autonomous Vehicles and Augmented Reality (AR)/Virtual Reality (VR)
IoT-enabled pressure sensors, integrated with Convolutional Neural Networks (CNN), provide real-time monitoring to identify potential bed sores. These sensors
continuously assess pressure points on a patient's body, and the CNN analyzes the data for early detection of risk. Through predictive analytics, caregivers receive
Identification and Prevention of Bed alerts, enabling timely interventions like repositioning to prevent pressure ulcers. This proactive approach enhances patient care and minimizes the occurrence of bed
CPG 7 sores/Pressure Ulcers using IoT sores by leveraging smart technology for constant surveillance.
Project Definition: The fasteners are usually mixed together after vehicle maintenance and need to be separated for future use
Solution Proposed: The goal of this project is to develop a hardware machine that can semi-automate the process of sorting nuts and bolts using machine learning. The
machine will be controlled by a microprocessor and will use a combination of hardware design, software programming, and machine learning techniques to identify
and sort nuts and bolts based on their characteristics.
CPG 8 A.S.S.I.S.T. Outcome: A Sorting machine
1. Problem Statement:
Inefficient parking management at Thapar College leads to
congestion and frustration among students and staff. Current
manual systems result in time wastage and contribute to
environmental issues. There's a pressing need for an intelligent
parking solution to optimize space utilization and enhance user
experience.
2. Objectives:
The key objectives include real-time slot recognition, vehicle
type classification, and providing users with recommendations
for alternative parking if the designated area is full. The system
will enhance efficiency, reduce congestion, and improve the
CPG 9 Thapar Smart Parking System overall parking experience for students and staff.
Glucose Measuring Device using hardware components – Optimal range, Glucose not in range (less or more)
Full Stack Website for Glucose Monitoring
Doctor and User Profile pages
Real time blogs can be added from Doctor’s profile
User can search Doctors
Doctor can view user’s details
ML based prediction based on data collected from user (and central data) whether patient is at risk or not; whether user should consult doctor or not
Real-time Data Visualization: Provide real-time graphs and charts to visually represent the user's glucose levels over time. This can help both users and doctors to
understand trends and make informed decisions.
Alerts and Notifications: an alert system to notify users and doctors about critical glucose levels or other important events. This could be through email, SMS, or in-app
notifications.
Cost effective
Machine learning-based calibration will be done to enhance the accuracy of glucose estimation
Non invasive method of glucose Calibration will be based on a deep neural network (DNN) model
CPG 10measuring using NIR spectroscopy. The validation involves testing on healthy, prediabetic, and diabetic samples, demonstrating its stability and accuracy.
 Capstone projeCts 2024

Project Definition: RapidBill Fusion will be an AI-powered autonomous shopping and checkout system for retail stores that combines the power of computer vision
and machine learning to provide an amazing shopping experience. It provides a faster checkout shopping experience to minimize human interactions.

Solution Proposed: We are developing an autonomous shopping checkout system that employs computer vision and machine learning technology to seamlessly process
purchased items. The system will utilize machine vision to identify and catalog items, automatically generating a detailed bill with the help of a pre-uploaded price list.
Users can conveniently complete the payment process through a user-friendly web interface, making the entire checkout experience efficient and hassle-free.

Outcome: The outcome will be a comprehensive billing and checkout system that leverages computer vision to identify items, generating accurate bills based on a pre-
uploaded price list. The system seamlessly integrates with a user-friendly web interface for convenient payment processing. Upon successful payment, users can
CPG 11RapidBill Fusion effortlessly collect their purchased items along with an electronic bill.
This project introduces an innovative solution to streamline supply chain management, combining the power of blockchain technology and digital twin integration. The
focus is on providing businesses with a user-friendly platform to track products, monitor manufacturing processes, and ensure end-to-end visibility.
Our website integrates a global supply chain map, pinpointing suppliers, manufacturers, and warehouses on a world map for streamlined visibility and management.
Digital Twin Integration in Supply We leverage blockchain for secure, decentralized record-keeping of orders and seamless data sharing with our digital twin platform.
Chain Management through We aim to create a digital twin for our manufacturing process, enabling real-time monitoring, fault detection, and proactive maintenance. This will optimize machine
CPG 12Blockchain Technology health, enhance efficiency, and provide valuable insights into our operations.
A system to digitize and centralize all hostel services to increase efficiency, save time, save money, and reduce paperwork thereby making the whole process more eco-
friendly.

Our solution consists of the following services:

- Hostel + room allocation system
- Cleaning management system
- Mess management system
- Laundry management system
- Complaint management system
- Discipline management system
- Leave management system
CPG 13Hostel Management System - Inventory management system
Objective: Develop a platform for college societies to streamline event information, registration, and management.
Features: Dedicated sections for each society, events, easy navigation, and a specialized section for Analytics dashboard for societies
CPG 14TIET Event Management System Goal: Enhance efficiency and engagement by consolidating all details on one website.
Overview:
StockPrecision Pro integrates advanced financial metrics, including Internal Rate of Return (IRR) and Weighted Average Cost of Capital (WACC), ensuring users
receive concise stock recommendations with vital financial insights. This evolution facilitates optimal investment decision-making. Additionally, the system now
incorporates candlestick chart analysis, providing users with visual insights into market trends for informed decision-making.
Benefits:
StockPrecision Pro empowers users with precise financial insights and visual trend analysis, facilitating optimal investment choices. The streamlined interface ensures
CPG 15StockPrecision Pro an efficient user experience, making it a valuable tool for strategic portfolio management.
Project Definition:
The project aims to develop an AI-driven comic generator designed to foster the appreciation of Indian traditional values. The system takes input in the form of
sentences or brief storylines and transforms them into visually appealing and culturally resonant comic illustrations.
Solution Proposed:
The AI comic generator employs advanced natural language processing (NLP) and computer vision techniques to interpret the input text and generate corresponding
illustrations.
Outcome:
CPG 16AI Comic Creator The envisioned outcome of this project is a revolutionary tool for the users to impart valuable cultural knowledge in an interactive and entertaining format.
Development of computer-aided and predict depression using EEG signals. It systematically reviews research conducted between 2016 and 2021, analyzing methodologies, algorithms, and outcomes.
depression diagnosis system using The study identifies patterns, evaluates approaches, and highlights future directions to improve the reliability of depression diagnosis through EEG-based deep learning
CPG 17wearable sensors and deep learning models.
network congestion or jammers disrupt conventional communication channels. This poses a significant challenge to effective coordination, potentially compromising
the safety and security of the environment.
Our project addresses this critical challenge of conventional communication networks becoming unreliable or congested. The project utilizes the LoRa chip, an IoT
technology, to enable individuals to form clusters for seamless communication without relying on internet, WiFi, or cellular connections. The initial deployment focuses
on enhancing communication among security guards during events and fests within a college campus.

Proposed Solution:
The project leverages the long-range capabilities and low-power consumption of the LoRa chip to provide a decentralized communication solution. Users can form
communication clusters dynamically, allowing for real-time, offline communication without relying on traditional networks. The system also incorporates location
tracking, enabling supervisors to monitor the real-time whereabouts of each guard. This innovative approach circumvents the limitations of internet-dependent devices.

Expected Outcomes:
Enhanced Communication: Security personnel can communicate seamlessly within formed clusters, ensuring efficient coordination even in areas with limited network
connectivity.

Improved Situational Awareness: Real-time location tracking enhances supervisors' ability to monitor and deploy personnel effectively, leading to improved situational
awareness during events.

Reduced Dependency on Traditional Methods: The system reduces reliance on walkie-talkies and traditional communication methods, providing a scalable solution
adaptable to various large-scale events and scenarios.

Other Relevant Details:

Scalability: As the project utilizes LoRa, which is a very versatile technology, it is easily scalable, making it suitable for deployment in diverse environments and
scenarios beyond the initial college implementation.

Low-Power Consumption: The LoRa chip's low-power consumption ensures prolonged battery life for devices, making it a sustainable and efficient solution for
extended deployments.
SecureMesh: Stay Secure, Stay
Connected. A LoRa based The project is poised to redefine communication standards for security personnel, offering a robust and reliable solution to overcome the challenges posed by limited
CPG 18communication & tracking solution network connectivity during critical events and gatherings.
Project Title: Facial Recognition with 3D Vision

Description:
Enhance facial recognition technology by integrating 3D vision. The project focuses on using advanced 3D depth-sensing to create detailed facial models, improving
accuracy, security, and adaptability. Key features include real-time recognition, anti-spoofing measures, cross-platform compatibility, and machine learning integration.
The system aims to find applications in secure access control, user authentication, attendance tracking, and more, providing a robust and efficient solution with
CPG 19Facial Recognition enhanced privacy measures.
We aim to develop an innovative deepfake detection project utilizing advanced machine learning algorithms to identify and mitigate the risks associated with
manipulated media content especially nowadays. The project will aim to research upon and then employ cutting-edge technology to analyze human features and detect
CPG 20DeepFake Detection anomalies, providing robust protection against the proliferation of deceptive multimedia in various online platforms.
Motor imagery EEG signal classification is a common task in brain-computer interface (BCI) systems and involves analysing electroencephalogram (EEG) data to
classify different motor activities. The goal is to develop a model that can distinguish between various motor tasks based on brainwave patterns captured by EEG and
then use that data to interpret brain signals such as to distinguish between imagining moving a hand or foot or actually doing so. The project aims to enhance
CPG 21EEG Motor Imagery understanding and application of EEG in motor imagery tasks.
 Capstone projeCts 2024

The objective of this project is to design and develop a technologically advanced smart glove capable of interpreting sign language gestures. The primary goal is to
enable seamless communication for individuals with hearing impairments by translating their sign language into audible speech. Additionally, the project aims to
GloveLingo: Empowering enhance inclusivity by incorporating a feature that converts spoken words from others into text, facilitating effective two-way communication between individuals with
Communication through Smart hearing impairments and those without. The ultimate aim is to create an assistive device that promotes accessible and barrier-free communication for the deaf and hard
CPG 22Gloves of hearing community.
Project definition:
Implementing a lid-based system for waste categorization can simplify the user experience and improve the accuracy of waste separation.
Solution Proposed:
Design a smart lid with integrated sensors, such as optical sensors or cameras, capable of identifying the type of waste placed on it.Calibrate the sensors to accurately
recognize different types of waste. Train the system to distinguish between recyclables, non-recyclables, and other waste categories.Integrate the sensors with the
microcontroller to process the data. Develop algorithms to interpret the sensor readings and categorize the waste accordingly. Provide feedback mechanisms such as
visual indicators or sound signals to inform users that their waste has been correctly categorized.
Outcome:
The proposed smart lid system has the potential to revolutionize waste management by providing a user-friendly, efficient, and accurate way to categorize waste,
CPG 23EnviroMate ultimately contributing to environmental sustainability and responsible waste disposal practices.
Problem:
Our system is supposed to monitor driving in real-time as well as creating a detailed log of each drive session .

Elaborating the problem:

Many a times it happens that driver is not the actual owner of the vehicle or driver is an employee in a transportation organisation.
So, the actual owner of the vehicle may not get to know how the driver actually drives the vehicle.

Solution:
1. Our system will determine location of vehicle through GPS and will determine the speed limit for that particular location and will keep a check on overspeeding in
real-time.
2. Using sensors we will create a Driver Attention Monitoring System which will check driver's attention on road while driving (i.e by detecting gestures of the driver).
3. The system will detect drifting, wrong overtaking and wrong turns etc.
4. The system will generate a log for whole drive session which will be accessible to the owner of the vehicle and will generate a warning notification in real-time in
CPG 24Drive Monitoring System case of adverse rash driving.
The purpose of Thapar Time Table Management System (TTTMS) is to create an essential system that will significantly improve the class schedules within TIET. The
objective of this system is to help in transiting from a manual, time-consuming process to an automated and efficient scheduling solution. The TTTMS will simplify the
creation and management of lecture/tutorial/laboratory schedules, reducing the workload on administrative staff and ensuring optimal resource utilization,
The TTTMS involves developing a dedicated software application that will consider course requirements, faculty availability, mom availability, and student preferences
Thapar Time Table Management to generate an optimized timetable. This system will also include features for automatic updates and notifications, making it easier to manage changes and substitutions
CPG 25System throughout the semester.
CPG 26Vehicle Tracking System using LoRa We are deploying a e-rickshaw tracking system using LoRa with a live dashboard.
In our advanced IoT-based Visitor Identification system, we have seamlessly integrated state-of-the-art liveness detection techniques to enhance security and thwart
potential spoofing attempts, including those involving deepfakes and 3D sculptures. When the controller detects a face using the HAAR algorithm, it not only activates
the camera to capture the incident but also employs a multi-modal liveness detection system. This system combines motion analysis and texture analysis techniques.
The integration of these techniques is adaptive, dynamically adjusting to environmental conditions and user behaviour, ensuring a comprehensive and resilient approach
to identity verification. In the event of a potential threat, the system delivers real-time notifications to the user via a messaging app, embedding a live photo of the
detected face for immediate validation on their Android phone. This sophisticated solution not only addresses the specific needs of patients with long-term memory loss
or Alzheimer's but also caters to elderly individuals who may not have the means to afford a caretaker, providing them with a reliable and secure visitor identification
system. The immediate threat detection, user-friendly notifications, and cost-effective care make it a practical and valuable solution for enhancing safety and security in
CPG 27Visitor Identification System real-time scenarios.
Breathaway, an innovative asthma management solution, prioritizes the well-being of children and seniors. The Android app provides real-time guidance, notifying
users when to press for medication and ensuring complete dose inhalation. Tailored for those who have recently been introduced to inhalers and may require spacers,
Breathaway records breaths per dose, total doses, and successful inhalations. It sends timely reminders and facilitates data sharing with healthcare professionals,
emphasizing the importance of effective asthma management for the vulnerable demographics. Breathaway aims to make asthma care accessible and efficient,
contributing to the health and safety of families, especially children and seniors. The Smart Inhaler for Children is an Arduino-based device equipped with pressure
sensors and Bluetooth connectivity, designed to actively monitor inhaler usage among school-going children. This innovative solution aims to enhance respiratory care
by tracking medication adherence and providing timely reminders. By leveraging real-time data, the device empowers parents and caregivers to proactively manage a
child's respiratory health, minimizing the risk of complications associated with conditions like asthma. The intelligent inhaler not only ensures accurate monitoring but
also acts as a digital companion, fostering consistent and effective medication routines through its built-in reminder system. Ultimately, this project strives to create a
CPG 28BREATHAWAY comprehensive and user-friendly tool that contributes to the overall well-being of children dealing with respiratory challenges.
The solar-powered agricultural rover project is an innovative initiative aiming to revolutionize precision farming. It features a solar-powered rover with a precision
spraying arm and soil sensors, enabling targeted resource application based on real-time soil conditions. Farmers can remotely control the rover through a mobile app,
monitoring sensor data for informed decision-making. The modular design encompasses key components like the chassis, mobility system, energy management, and
more. Environmental benefits include reduced resource wastage, economically leading to increased crop yield and reduced costs. The mobile app interface attracts
younger farmers to engage in precision agriculture, highlighting the project's potential to reshape traditional farming for sustainability and efficiency. Future work may
involve AI integration, multi-terrain adaptability, and collaborative farming, marking a significant step toward a more precise, environmentally conscious, and
CPG 29AgroRover economically viable agriculture.
Ensuring a Secure Operating The project aims to enhance the safety and performance of electronic vehicle (EV) batteries by implementing a comprehensive system that utilizes digital clone
Environment for Electronic Vehicles technology, cloud communication, and various sensors for real-time monitoring. The primary objective is to create a secure operating environment that maximizes the
CPG 30Batteries efficiency and longevity of EV batteries while ensuring the safety of both the vehicle and its occupants.
SentinelEye reimagines security through a unique blend of cutting-edge technology and adaptability. This "plug-and-play" system transcends reactive surveillance,
offering real-time monitoring and proactive threat detection. SentinelEye leverages the power of AI, specifically enhanced computer vision and object detection, to
identify a range of violations, from smoking and fire outbreaks to trespassing and weapons possession, in various environments. Its effectiveness hinges on the
integration of drone technology, enabling comprehensive coverage of large areas from elevated vantage points. This eliminates dependence on stationary cameras,
enhancing operational efficiency. To ensure precise image recognition, SentinelEye employs advanced image improvement techniques like ESRGAN. Furthermore, the
system provides real-time monitoring and rapid response capabilities through its user-friendly Software as a Service (SAAS) application. Preloaded models and
customizable features allow SentinelEye to seamlessly adapt to diverse settings, from academic institutions and shopping centers to construction sites and hazardous
waste management facilities. This versatility, coupled with robust cybersecurity safeguards, positions SentinelEye as a transformative force in global security, with the
CPG 31SentinelEye potential to Fins
Definition: significantly
& Ferns impact
presentssafety across various
a revolutionary sectors.aquaponics system that optimizes plant and fish growth while minimizing human intervention, Leveraging
automated
sensor technology and intelligent control systems. Fins & Ferns fosters a thriving symbiosis between plants and fishes, maximizing efficiency and sustainability.
Proposed Solution:
Environmental Optimization: Precise sensors and actuators maintain ideal temperature, humidity, and light conditions, ensuring optimal growth for both plants and fish,
regardless of external fluctuations.
Automated Water Management: Smart sensors detect low water levels and automatically replenish the system from a dedicated buffer tank, guaranteeing system
stability and eliminating manual intervention.
Real-time pH Monitoring and Adjustment: An integrated system continuously monitors and adjusts the water's pH balance, creating a healthy and thriving environment
for both aquatic and terrestrial life.
Outcomes:
Enhanced Yield and Growth: Optimized environmental conditions and automated water management lead to increased plant and fish production, maximizing yield and
resource utilization.
Reduced Operational Costs: Automation eliminates the need for manual adjustments, saving time and labor, while cloud-based monitoring allows for proactive
maintenance and minimizes downtime.
Sustainable Practices: Closed-loop water management, reduced waste generation, and efficient resource utilization promote environmentally responsible food
CPG 32Fins & Ferns production.
Graph Attention Networks for Traffic
CPG 33Flow Prediction Using the traditional time series prediction models we will construct graph attention networks for traffic flow prediction.
 Capstone projeCts 2024

LipSyncX, at the forefront of AI-powered lip reading, employs sophisticated deep learning and computer vision algorithms for precise decoding of spoken language
through meticulous analysis of lip movements. The system's robust architecture, trained on extensive datasets, guarantees good accuracy. Demonstrating a commitment
to continuous improvement, our team is actively exploring advancements, with future plans to integrate Large Language Models (LLMs) to further enhance accuracy
and linguistic nuances. This technical innovation aligns LipSyncX with industry-leading standards in AI-powered language processing, ensuring a seamless and reliable
communication tool for a diverse user base.

Beyond its transformative social impact, LipSyncX plays a crucial role in public safety and security. By empowering individuals with hearing disabilities, LipSyncX
fosters inclusivity, enabling more active participation in social interactions. Simultaneously, its advanced lip reading capabilities have significant applications in
enhancing strategic intelligence and law enforcement efforts. Whether in emergency situations, crowd monitoring, or criminal investigations, LipSyncX emerges as an
CPG 34LipSyncX impactful and versatile solution, contributing to both societal well-being and national security.
CaptainFarm's project develops a holistic defence system for farms. It integrates advanced surveillance for real-time animal detection, comprehensive identification,
and deterrent mechanisms to protect crops and livestock. The system also includes fire and soil quality sensors, along with a smart sprinkler system for optimised water
usage. Expected outcomes include reduced crop vandalism, enhanced protection, early fire detection, optimised water usage, and increased farm efficiency. The
CPG 35CaptainFarm project empowers farmers with proactive insights for sustainable practices.
Pothole Detection using conditions or some other bad weather conditions. Crowdsourcing can help by allowing drivers to report potholes and alert others in real-time, making roads
CPG 36Crowdsourcing safer for everyone
We’re developing an intelligent light management system featuring real-time occupancy detection using camera analytics, ensuring prompt illumination deactivation
AmbiSync: Room Atmosphere meets when no individuals are present. Additionally, our temperature sensing management system employs dynamic air conditioning control, responding adaptively to the
CPG 37Automation ambient temperature for optimised energy efficiency and occupant comfort.

Indoor Positioning System: An Indoor Positioning System (IPS) can be used to locate shoppers and guide them to their destination. IPS uses Wi-Fi, Bluetooth, and
other wireless technologies to determine a shopper's location and provide turn-by-turn directions.Before making a purchase, leading to a better decision-making
process.
Overall, a mall and shop navigation feature can enhance the shopping experience for customers, making it easier and more enjoyable to navigate through a large mall or
store.

Integrating a feature to add markers on stores with ongoing offers in the Indoor Positioning System (IPS) can significantly enhance the shopping experience. This
allows customers to not only navigate through the mall but also discover and take advantage of current promotions, contributing to a more informed and rewarding
shopping journey.

AR (Augmented Reality) based shopping features can enhance the shopping experience for users in a web application. With AR, users can visualize products in a more
realistic way The mall and shop navigation feature enables customers to find their way around the mall and easily locate their favorite stores as well as different sections
of each store.

AR Try-on: AR can allow users to virtually try on clothes or accessories in real-time. This can help users to find the right size and fit, which can be challenging when
shopping online.

The online billing and product scanning feature is integrated with a barcode scanner, making it easy for customers to scan and pay for their products without having to
stand in long queues.

The booking online tickets for fun zones in the mall feature enables customers to pre-book tickets for entertainment zones in the mall, avoiding long queues and
ensuring a hassle-free experience.

A security feature that generates a QR code after purchase provides an additional layer of security, ensuring that all purchases are confirmed and authorized before
customers leave the mall. This helps to prevent theft and fraud, providing customers with peace of mind as they shop.
The web application also allows customers to order food from the food court in the mall, with a seamless integration with the mall's POS system. Customers can select
from a variety of options and pay online, reducing the waiting time at the food court.
CPG 38Nav & Shop
Developing a robust deep learning model capable of translating images of food into detail and accurate recipe, an enhanced model which will help to recognise
multiple ingredient in single image.
The outcome for this project will be a user friendly application that allow user to translate food image into detail recipe, complete with list of ingredients. The model
CPG 39Image to Recipe Translator will be accurate and adaptable to various cuisines, valuable tool for individuals looking for creative recipe ideas
CPG 40Deep Fake Detector Outcome- Upon file upload the system employs advanced algorithm to analyse the content and determine whether it exhibits characteristics of deep fake manipulation
Accessing quality healthcare services and obtaining prescribed medications efficiently can be challenging, especially during times of pandemics. Traditional pharmacy
systems are often time-consuming, and finding the right doctor or scheduling appointments can be cumbersome. Additionally, the manual process of obtaining
prescribed medications from pharmacies may pose challenges during emergencies.

Proposed Solution:
Our project, is an online healthcare platform designed to streamline the entire healthcare process. It includes a comprehensive doctor recommendation system,
appointment scheduling, and an innovative medication dispensing solution. Users can search for suitable doctors in their area, book appointments, and receive updates
on prescribed medications through the platform.

Upon the completion of a medical checkup, doctors will update the prescribed medications on the website. Each medication is assigned a unique ID that is used in
generating a QR code, which serves as input for our hardware prototype - a medication dispensing vending machine.

The dispensing machine is a cutting-edge solution that acts as a pharmacy substitute. Users can scan the QR code generated for their prescribed medications, and the
machine dispenses the required medicines promptly. These vending machines can be strategically installed in hospitals, clinics, and public spaces, offering a convenient
and contactless method of medication retrieval.

Expected Outcomes:

Improved Healthcare Access: The project enhances accessibility to healthcare services by simplifying the process of finding suitable doctors and scheduling
appointments online.

Efficient Medication Dispensing: The vending machine prototype ensures a quick and hassle-free process for obtaining prescribed medications, reducing the burden on
traditional pharmacies.

Contactless Healthcare Solutions: Especially valuable during pandemics, the idea minimizes physical contact in the healthcare process, promoting a safer environment
for both patients and healthcare providers.

Optimized Resource Utilization: By automating medication dispensing through vending machines, the project contributes to the optimization of healthcare resources
Revolutionizing Healthcare Access and reduces the dependency on human-run pharmacies.
CPG 41and Medication Dispensing
Project Definition:
The project aims to design a Smart Weather-Responsive Irrigation System that utilizes real-time weather data to automate and optimize irrigation schedules and water
usage for agricultural purposes. By integrating meteorological information with soil moisture levels, the system seeks to provide a sustainable solution to the challenge
of water conservation in farming.
Proposed Solution:
The proposed solution is to design an advanced algorithmic software that processes real-time weather data, using sensors and weather integration for precise crop
hydration, which control water levels and optimize yields, sourced from public APIs, and correlates it with crop water requirements to generate dynamic irrigation
schedules. This software will include features for manual user inputs, allowing customization based on specific crop types and local soil conditions. It will also provide
a user-friendly interface for remote monitoring and manual override capabilities, ensuring flexibility and control remain with the farmers.
Expected Outcome:
The implementation of this smart irrigation system is anticipated to significantly reduce water consumption by tailoring watering schedules to actual environmental
needs, thus conserving a vital resource. Additionally, it is expected to enhance crop yields by providing precise watering, improve the efficiency of agricultural
CPG 42AgriAid: Smart Irrigation System operations, and offer farmers greater control over their irrigation practices through advanced data analytics.
 Capstone projeCts 2024

A Non-Invasive Approach to
Amputee Rehabilitation: Decoding Final description: Research would focus on a cost-effective upper limb prosthesis for amputee, integrating deep learning techniques and multiple sensors. The
EEG Signals Using Advanced Signal prosthesis,aims to provide enhanced functionality. Data from these sensors form a comprehensive dataset for training, optimizing model performance with limited data.
CPG 43Processing Techniques The project would address challenges faced by upper limb amputees, offering a sophisticated and affordable solution for rehabilitation.
Problem Definition:
Urban parking challenges require an innovative approach due to space scarcity, inefficiencies, and manual management pitfalls.

Proposed Solution:
Introducing an Intelligent Parking Management System (IPMS) using OCR technology for swift license plate recognition and automated space allocation.

Key Features:
1. OCR Technology: Accurate license plate recognition.
2. Real-time Data Integration: Dynamic parking space allocation for optimal utilization.
3. User-Friendly Interface: Intuitive application for seamless bookings.
4. Notification System: Instant alerts for parking status updates.
5. Secure Transactions: Trustworthy and transparent reservations.

Expected Outcomes:
1. Efficient Parking Utilization: Minimizing congestion and improving traffic flow.
2. Reduced Manual Workload: Automation for error reduction and operational efficiency.
3. Enhanced User Experience: A user-friendly interface for convenient parking.
4. Improved Enforcement: Efficient monitoring and reporting for rule enforcement.
5. Data-Driven Insights: Valuable data for future urban planning.

Conclusion:
CPG 44ParkEasy The Smart Parking Solution, integrating OCR technology, aims to revolutionize urban parking, enhancing efficiency and transparency for a seamless user experience.
Medical Image Super-Resolution Medical Image Super-Resolution using Generative Adversarial Networks (GANs) is an approach to improving the resolution and quality of medical images. The goal
Using Generative Adversarial is to enhance the resolution of low-quality medical images (e.g., MRI, CT, X-ray, ultrasound) while preserving crucial diagnostic details. GANs achieve this by
CPG 45Networks generating high-resolution images from their low-resolution counterparts.
We would like to develop a wearable prototype combining EEG and eye tracking to detect human emotions in real-time.
As background, We recently came across an idea about "EmotionMeter: A Multimodal Framework for Recognizing Human Emotions" that proposes a novel six
electrode EEG setup with eye tracking for emotion detection. The key benefits are good accuracy in classifying four emotions, complementary information from the
two modalities, and reasonable stability over time demonstrated through cross-session evaluations.
For our project, we plan to recreate this setup using low-cost off-the-shelf components and test it ourselves before conducting trials with more participants. The goal is
CPG 46Multimodel Emotion Recognition to validate the findings of the paper and demonstrate a practical wearable affective computing system.
Problem Definition:
The problem at hand is the need for a temperature-controlled glove to address challenges associated with maintaining optimal hand temperature in cold environments.
Extreme temperatures can lead to discomfort, reduced dexterity, and potential health issues. This is particularly relevant for individuals working in industries such as
construction, outdoor activities, or those facing extreme cold weather conditions.

Proposed Solution:
Our proposed solution is a technologically advanced temperature-controlled glove that incorporates heating elements to maintain a comfortable and consistent
temperature for the wearer's hands. The glove will be equipped with sensors to monitor the ambient temperature and the user's hand temperature, allowing it to adjust
the heating elements accordingly. The user will have the option to manually control the temperature settings based on their preference.

Expected Outcomes:

1. Enhanced Comfort: Users will experience improved comfort, especially in extreme temperatures, leading to increased productivity and overall well-being.

2. Increased Safety: Maintaining optimal hand temperature is crucial for preventing conditions like frostbite, thereby enhancing user safety.

3. Versatility: The temperature-controlled glove will be versatile enough to be used in various industries and activities where hand protection and temperature
CPG 47Thermal Glove regulation are essential.
Project Definition:
The project aims to address the mobility challenges faced by individuals affected by conditions like Guillain-Barré Syndrome (GBS). The primary issue is the
compromised functionality of the limbs, necessitating the exploration of an alternative means of movement and independence.

Solution Proposed:
To enhance autonomy for patients with impaired limb function, we propose the development of a wheelchair driven by upper-body movements above the neck, such as
eye and neck motions. This innovative approach seeks to provide individuals with the ability to navigate and perform daily activities independently, reducing their
dependence on external assistance.

Outcome:
The envisioned outcome of this project is a technologically advanced wheelchair that translates precise above-the-neck body movements into smooth and responsive
Smart Wheelchair for Guillian Barre mobility. This solution aims to empower individuals affected by GBS, enabling them to regain a sense of control over their movements and foster greater independence
CPG 48Syndrome(GBS) and similar diseases in their daily lives.
Some Key features are:
1. Emergency Signal and Medical Information Transmission in case a crash is detected by the device:
• Automatic transmission of an emergency signal to the nearest medical facilities .
• Transmission of coordinates and essential rider information for quick medical assistance.
2. Family Notification System:
Instant dispatch of an emergency SMS to pre-defined family members upon a crash.
Ensures timely communication with family, keeping them informed about the rider's situation.3.
3. Trip Monitoring and Report (mainly for parental control purposes):
Utilization of GPS technology to record and store trip data on web/website, including average and top speed, along with the route. User-friendly interface for accessing
detailed trip histories with other useful information promoting transparency and accountability .
4. Speed Limit Alert System:
Generation of a distinctive beeping sound to alert the rider when surpassing a preset speed limit.
Smart IoT based helmet device for Encourages safer riding practices by reinforcing awareness and adherence to speed limits.
enhanced rider safety and
CPG 49connectivity This project idea is perfect for hardware integration with software to solve a real problem that exists.
Drug discovery is a long and complex process that requires a significant amount of
investment in terms of time and resources.
The development of cancer drugs is a particularly challenging area of drug discovery, as it requires the identification of compounds that selectively target cancer cells
while sparing normal cells.

In this project we focus on the use of machine learning to develop predictive models for drug solubility. To enhance our understanding of compounds, we plan to
Drug discovery for Cancer using Data extract biological activity data from the ChEMBL database. This information can be used to develop QSAR models that can predict the activity of compounds against
CPG 50Analysis and ML various biological targets.
website. Harnessing the dataset from the Indian Sign Language Research and Training Center (ISLRTC), our innovative solution combines the power of vision
processing, sequential learning, and pose estimation. It also enhances accessibility and communication for the hearing-impaired, fostering inclusivity in diverse
CPG 51Sign Language Interpretor communities.
Study Buddy is an AI and AR-powered educational platform designed to enhance the learning experience of pre-primary children through dynamic, multisensory
activities. By integrating
advanced technologies such as artificial intelligence, augmented reality, and voice recognition, StudyBuddy creates an immersive and engaging environment that caters
CPG 52Study buddy to diverse learning needs.
 Capstone projeCts 2024

I am Akshat Kumar Jain from the Computer Engineering (COE) branch. I am reaching out to express my and my team's keen interest in undertaking the capstone
project under your guidance and extending the work you have previously conducted which aims to explore the effects of yoga on brain activity and cognition by
employing Electroencephalography (EEG) cap technology. Our team's proposed capstone project would aim to -
1. Include the factor of the multimodal (EEG, ECG, EMG) approach in analyzing the effects of yoga on brain activity.
2. We will try to increase the accuracy of the model.
Effects of Yoga on Brain Activity and 3. We will also try to not only analyse the effect of yoga but also the effect of different genres of music on our brain activity.
Cognition by Employing
Electroencephalography (EEG) Cap If you find our proposal aligning with your current research goals, we would be honored to discuss this further at your earliest convenience.
CPG 53Technology We appreciate your consideration of our proposal and are enthusiastic about the prospect of working under your guidance.
A website where you can just enter your budget , which type of place you wanna go(mountains ,beach, desert etc) and the number of days for vacation. Then we use an
AI based model to create a whole package under the budget checking with other sites to book transport to hotel and all other amenities . The AI/ML model also
searches and makes decision based on traffic at the current place and suggests places with less traffic and same beauty which are underrated. It also suggests best
places based on weather forecast. we also implement a chat bot which translates language for the user on the website and as well as in real time for convenience in
CPG 54Desti-Ease conversing to the locals and understanding the local language.
A simple-to-use file-sharing system that utilizes socket programming to break and distribute a file across a network and allow its users to download those files easily in
Smart Parking Guidance System a faster and more efficient way. We plan to extend the capabilities of this project bsfers become mory introducing compression algorithms and other features to ensure
CPG 55using IoT and Machine Learning that file trane efficient and secure as compared to our previous iterations.
Graph Neural Network and
Tranformer-based Chemical Graph Neural Network and Tranformer-based Chemical Reaction outcome generator
CPG 56Reaction outcome generator. Using graph neural networks we try to find the outcome of a chemical reaction.
ECG signals play a critical role in diagnosing various cardiovascular conditions, yet collecting
labeled data for training robust machine learning models remains a challenge due to limited
availability and privacy concerns. Traditional data augmentation techniques, such as random noise
addition or time shifting, often fail to capture the complexity and variability present in ECG data,
leading to suboptimal model performance. The aim of this capstone project is to develop and
Time Series Based Generative evaluate a novel approach utilizing time series Generative Adversarial Networks (GANs) for data
CPG 57Adversarial Network augmentation of ECG signals.
Project Definition:
Our community skill exchange platform aims to foster collaboration and knowledge-sharing by facilitating non-monetary transactions between members. By leveraging
the diverse skills and talents within our community, individuals can offer their expertise in various fields in exchange for the services or learning opportunities they seek.
Solution Proposed:
The platform will feature user profiles where members can showcase their skills, expertise, and the services they're seeking in return. Users can browse through listings,
and search for specific skills or services.
More intricate features are upcoming for this.
Outcome:
By establishing a community-driven skill exchange platform, we anticipate several positive outcomes. Firstly, members will have access to a diverse pool of skills and
resources without the barrier of monetary transactions. This promotes inclusivity and equal participation within the community. Secondly, individuals can expand their
skill sets, broaden their networks, and foster meaningful connections with fellow members. Lastly, by promoting collaboration and knowledge-sharing, the platform
CPG 58Community Skill Exchange Platform contributes to the overall growth and development of our community, creating a supportive environment for lifelong learning and skill enhancement.
IoT-Enabled Soil Nutrient Analysis we propose a new IoT-enabled soil nutrient classification and crop recommendation (IoTSNA-CR) model to classify soil nutrients and recommend crops for precision
and Crop Recommendation Model agriculture. The IoTSNA-CR model involves different processes, namely data acquisition using sensors, storage in the cloud, MSVM-based classification with FFO-
CPG 59for Precision Agriculture based parameter optimisation, and crop recommendation
Objectives:
1. 1. Explore and implement advanced image processing techniques to extract meaningful features from the image data, considering the multi-filter representation of
each sample.
2. Develop a deep learning-based classification framework capable of accurately predicting the subcellular localization labels of proteins within human cells.
Huma Protein Atlas Image 3. Evaluate the performance of the proposed classification system extensively using appropriate metrics and validation methodologies, considering the diversity of cell
CPG 60Classification types represented in the dataset.
Autonomous Recommendation- Imagine strolling through the grocery store with a helpful companion that not only keeps track of your items but also calculates the total price as you go. This is the
CPG 61based Intelligent Cart magic of the Smart Cart, a revolutionary shopping innovation that streamlines your shopping experience.
The "3D Virtual Try-on" project aims to revolutionize the fashion industry by seamlessly integrating augmented reality (AR) technology into the digital fashion
landscape. In response to the rapid evolution of technology, this project addresses the challenge of distinguishing between digital and physical clothing. The proposed
solution involves the development of a proprietary AR clothing try-on technology, incorporating advanced features such as 3D body tracking, real-time multi-class
segmentation, and cloth simulation. The innovative technology allows users to virtually try on digital garments, bringing a level of utility and realism comparable to
physical pieces, all without the need for physical production. The anticipated outcome of this project is a paradigm shift in the way consumers interact with fashion in
the digital realm, transcending traditional boundaries and offering a dynamic and immersive virtual try-on experience. This initiative not only propels digital fashion
CPG 623D Virtual Try-on from the virtual world into our daily lives but also sets the stage for the future of fashion by redefining the way we engage with clothing in the digital age.
This project implements a markerless motion capture system to enable real-time 3D pose tracking within the Unity3D environment. By leveraging advanced pose
estimation techniques, the system extracts human motion data directly from a standard camera feed and maps it to a 3D character model, UnityChan, for seamless
animations.

The key functionality revolves around capturing live pose keypoints, processing the data for accuracy, and visualizing the motion through responsive 3D character
animations. This framework eliminates the need for expensive motion capture hardware or markers, making it accessible and cost-effective.

Key features include:

Real-Time Animation: The system ensures low-latency pose tracking for smooth and realistic animations.
Markerless Technology: Simplifies the setup process by using standard cameras instead of specialized equipment.
Unity3D Integration: Utilizes Unity’s robust rendering and animation tools to create visually appealing outputs.
Real-Time 3D Pose Tracking Using This project has applications in various domains, including gaming, VR/AR experiences, healthcare for motion analysis, and animation production, showcasing its
CPG 63Unity3D and UnityChan potential as a flexible and scalable solution for real-time motion capture.
This capstone project involves the design and development of an autonomous rover equipped with a 76-degree rotating camera, aimed at real-time environmental
mapping and localization. The rover is controlled via a Flutter-based Android application that manages the movement of the rover, its tires, and the camera's rotation.
The camera captures multiple images as it rotates, which are then processed and stitched together to create a comprehensive 3D map of the surroundings. This 3D map
is integrated into a virtual reality (VR) environment using Unity, allowing users to explore the mapped area in an immersive 360-degree view through a VR headset.
The project addresses several challenges in autonomous navigation, real-time data processing, and VR integration, offering significant advancements over existing
solutions that rely on single-image localization or manual control. The methodology includes a thorough literature review, stakeholder engagement, iterative
prototyping, and testing, as well as the comparative analysis of tools and technologies. The result is a robust and versatile system that enhances safety, operational
efficiency, and user experience across applications such as search and rescue, environmental monitoring, and industrial inspections.
This project demonstrates the potential of integrating robotics, computer vision, and VR technologies to create a sophisticated autonomous system capable of real-time
CPG 64Immersive Reconnaissance Mapping mapping and immersive visualization, paving the way for future research and development in autonomous systems.
 Capstone projeCts 2024

development of targeted treatment strategies. Here are some key reasons why image fusion is important in Alzheimer's disease:

1. Comprehensive Assessment: Alzheimer's disease is a complex neurodegenerative disorder that affects both the structure and function of the brain. Image fusion
allows for a comprehensive assessment by combining information from different imaging modalities, such as structural MRI, functional MRI, PET, and SPECT. This
comprehensive approach provides a more detailed and nuanced view of the brain's changes in Alzheimer's patients.

2. Early Detection and Diagnosis: Early detection of Alzheimer's disease is crucial for timely intervention and treatment. Image fusion helps in identifying subtle
structural and functional changes in the brain that may occur in the early stages of the disease. Combining multiple imaging modalities can improve the sensitivity and
specificity of diagnostic tools, enabling healthcare professionals to detect Alzheimer's disease at an earlier and potentially more treatable stage.

3. Understanding Disease Progression: Alzheimer's disease progresses over time, leading to changes in both brain structure and function. Image fusion allows
researchers and clinicians to track these changes longitudinally. By combining data from different time points, they can gain insights into the progression of the disease,
identify biomarkers, and evaluate the effectiveness of interventions.

4. Differentiating Alzheimer's from Other Conditions: Many neurodegenerative disorders share common symptoms, making accurate diagnosis challenging. Image
fusion helps in differentiating Alzheimer's disease from other conditions by providing a more comprehensive and specific characterization of brain abnormalities. This
aids in developing targeted treatment plans and improving patient outcomes.

5. Research and Treatment Development: Image fusion is instrumental in advancing research on Alzheimer's disease. Researchers can use fused imaging data to
identify specific patterns, correlate structural and functional changes, and uncover underlying mechanisms of the disease. This knowledge is critical for developing
targeted therapies and interventions.

6. Personalized Medicine: Alzheimer's disease exhibits considerable variability among individuals. Image fusion can contribute to the development of personalized
medicine approaches by tailoring treatments based on an individual's unique brain characteristics. This personalized approach may lead to more effective and targeted
interventions.

7. Validation of Biomarkers: The identification and validation of reliable biomarkers are essential for Alzheimer's research. Image fusion can help validate potential
biomarkers by correlating structural and functional changes with disease progression. Validated biomarkers are crucial for developing diagnostic tools and monitoring
Topic :Alzheimer's disease Image treatment response.
CPG 65fusion
The objectives of this project are multifaceted, covering various aspects such as technological development, user engagement and many more. The prime objectives
that will be addressed are mentioned below:
1) Create accurate and reliable algorithms for live yoga pose detection.
2) Ensure the system can recognize a variety of yoga poses in real time.
3) To develop algorithms that can provide constructive and personalized feedback to users for improving their yoga poses.
4) Ensure the correction system is adaptable to different skill levels and body types.
CPG 66Yoga Pose Detection and Correction 5) Achieve a high level of user engagement by creating an app that is user-friendly and enjoyable to use.
Machine Learning based Image
CPG 67Dehazing
The development of deepfake technology has raised serious concerns about the veracity and authenticity of digital material, which has broad ramifications for a number
of industries, including security, entertainment, and journalism. In order to overcome these obstacles, the DeepTrace: Deepfake Detection System project focuses on
creating a reliable detection system that is capable of precisely identifying and analyzing manipulated media content. The system uses state-of-the-art machine learning
methods to identify deepfake frames in digital video, namely Convolutional Neural Networks (CNNs) and Generative Adversarial Networks (GANs). DeepTrace
works in stages. The first step is identifying deepfake frames by looking for patterns and other characteristics that are frequently found in deepfakes. Subsequently, the
DeepTrace: Deepfake Detection system retrieves the indicated frames for more thorough examination and confirmation. In the last phase, the degree of modification is assessed by comparing the
CPG 68System (DDS) extracted frames to the original media, which sheds light on the techniques employed to create the deepfake.
Our project focuses on creating a robust system capable of transforming real-time streamed foggy images into clear, high-quality visuals using advanced deep learning
techniques and sophisticated architectures. This innovative solution aims to serve as a foundational technology for a range of automated and semi-automated
applications, specifically designed to assist both manually driven and self-driving cars in navigating through challenging and foggy weather conditions. Our goal is to
contribute to the development of a versatile, AI-driven tool that ensures safety and efficiency in adverse weather scenarios. This project not only addresses immediate
Deep Learning-Based Defogging for challenges in transportation but also lays the groundwork for transformative advancements in various domains reliant on clear, accurate visual data in the face of harsh
Enhanced Image Clarity in weather conditions. Employing advanced deep learning, our system is tailored for real-time defogging of streamed images, enabling enhanced drone surveillance
CPG 69Challenging Environments capabilities, efficient traffic control, and improved navigation for both manual and self-driving vehicles in harsh and foggy conditions.
Problem: Many people enjoy having gardens but don't take care of it for long because of their busy schedules resulting in death of number of plants due to lack of
maintenance.

Solution: The main aim of this project is to provide a solution by developing IoT-based platform along with SprinklerBot and Buggy. We aim to develop a smart
garden system which has smart cameras that identify plant types for ensuring precise watering (automating the irrigation process by analyzing the moisture of soil and
the climate condition like raining) and fertilization. A system that can perform auto-monitoring of humidity and temperature of surroundings, air quality and soil
moisture. Furthermore, all the notifications would be sent on the application about the tasks performed by the system.

GardenMate- Automated Home Expected Outcomes: We look forward to design an automatic system which keeps the track of garden environment and acts according to the requirements(like water
CPG 70Garden Care System and fertilizers). Also, one should receive notifications daily keeping them informed about their garden's well being.
In this project we will develop a software for tracking inventory levels,orders,sales and deliveries and we can create a work order,bill of order and other production
CPG 71Inventory management system related documents.
This Mobile App will train a Machine Learning model based on user's preferences of images he/she wants to keep/delete. Then it will provide the opportunity to delete
CPG 72Delete It all images at once based on user's preferences.
1. Introduction:
The Smart Gloves Sign Language Translator is an innovative solution designed to bridge the communication gap between individuals who use sign language and those
who may not be proficient in it. This project aims to provide real-time translation of sign language gestures into spoken language, fostering better communication and
understanding in various social contexts.

2. Solution Proposed:
The core of the project involves the development of smart gloves equipped with sensors and a communication module. These gloves are designed to capture and
interpret sign language gestures made by the user. The system utilizes machine learning algorithms to recognize and translate these gestures into text or spoken
language.
3. Outcome:
Improved Communication: The Smart Gloves Sign Language Translator enhances communication between individuals who use sign language and those who do not,
fostering inclusivity and understanding.

Increased Independence: Users of sign language gain increased independence in various scenarios, such as ordering food, asking for directions, or participating in
group discussions.

CPG 73Smart gloves sign language translator Educational Tool: The system serves as an educational tool, helping individuals learn sign language by providing real-time feedback on their gestures.
Our project focuses on changing how emergency services and medical care are provided by making it easy and safe for people to access their medical histories during
crashes and other serious situations.

We propose to make a personalized medical app that includes:

1) An SOS feature that can help medical personnel access users' medical information.
2)User's medical details, allergies, medical reports, and doctor's prescriptions.
3) A personalized AI/ML model that suggests medicines and remedies based on the user's medical reports and prescriptions through federated learning that enhances
CPG 74WellNex: Emergency Medical Portal the user's privacy.
The Deepfake Detection project aims to create a system capable of identifying manipulated content generated through deep learning techniques.By using smart
algorithms and machine learning, we aim to teach the system to tell the difference between real and fake content. The goal is to make the system accurate and strong
CPG 75Deepfake Detection enough to tackle the growing challenge of deceptive media, helping to prevent the spread of false information and scams.
 Capstone projeCts 2024

Problem Definition:
Thapar students often struggle to find and navigate scattered information crucial to their academic and college life. Timetables are confusing, admissions processes lack
clarity, course materials are fragmented, placement trends are opaque, and information about clubs, events, and faculty is dispersed across various platforms. This
creates frustration, hinders efficiency, and limits student engagement.

Proposed Solution:
ThaparGenie serves as students' premier conversational AI assistant, offering a centralized information hub leveraging natural language processing and machine
learning capabilities. This innovative platform enables students to access a wealth of resources and support seamlessly. With ThaparGenie, students can effortlessly
navigate vast chunks of information that would otherwise be scattered across our institute's various websites, documents and communication emails. Everything from
timetables and course materials to placement trends and admission FAQs will instantly be made available to the users as an answer to a simple prompt. Furthermore,
ThaparGenie simplifies locating faculty members, offering their contact information, expertise, and research interests at students' fingertips. Additionally, ThaparGenie
would provide information related to extracurricular events being conducted on campus and insights into the various student chapters and clubs. With its user-friendly
interface and extensive features, ThaparGenie revolutionizes student experience and information accessibility at the Thapar Institute of Engineering Technology
campus.

Expected Outcomes:
ThaparGenie is anticipated to deliver several key outcomes for the Thapar Institute of Engineering Technology community. By improving information access, students
will enhance academic performance and efficiency. The clarified admissions process will empower prospective students to make more informed decisions. Enhanced
ThaparGenie - “Where Queries Meet campus engagement will foster a more active student community, while streamlined communication channels will facilitate collaboration between students and faculty,
CPG 76Clarity” enriching the overall educational experience.
Definition:
Advanced Nocturnal Security with Concealed Multispectral Impostor Detection, Recognition, and Path Tracking using IR Cameras is a security system designed for
low-light conditions. It employs concealed infrared cameras operating in multiple spectral bands, advanced algorithms, facial recognition, and path tracking to covertly
identify and track potential impostors.

Proposed Solution:
Concealed IR Cameras: Covert surveillance with infrared cameras capturing multispectral data.
Impostor Detection Algorithm: Advanced algorithms analyze spectral signatures and movement patterns for distinguishing impostors.
Recognition System: Integrates facial recognition and biometrics for accurate identification.
Path Tracking: AI-powered trajectory analysis monitors and predicts the movement of individuals.
Real-time Alerts: Instant notifications to security personnel based on predefined criteria.

Outcome:
Enhanced Security: Robust defense during low-light hours against potential threats.
Advanced Nocturnal Security: Covert Surveillance: Discreet operations to minimize detection and evasion.
Concealed Multispectral Impostor Accurate Identification: Improved accuracy in identifying individuals, reducing false positives.
Detection, Recognition, and Path Proactive Response: Swift action enabled by path tracking and real-time alerts.
CPG 77Tracking with IR Cameras Adaptability: Versatile deployment in diverse security settings for heightened protection.
NeuralNetIoT is an advanced project aimed at refining IoT traffic analysis and classification using cutting-edge deep learning techniques. It addresses the need for
precise categorization of IoT data streams to enable efficient network management and enhanced security measures.
Solution Proposed:
NeuralNetIoT proposes a comprehensive approach to achieve its objectives:

1. **Data Collection and Preprocessing:**

- Raw IoT data streams are collected from various devices and sensors within the network infrastructure.
- Preprocessing techniques are employed to extract pertinent features and organize the data into standardized packets for further analysis.

2. **Deep Learning Model Implementation:**

- State-of-the-art deep learning models, including Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), are utilized to analyze packet-
level data.
- These models are trained on labeled datasets to discern intricate patterns and correlations within IoT traffic.

3. **Multi-Class Traffic Classification:**

- NeuralNetIoT classifies IoT traffic into a diverse range of categories to provide comprehensive insights for network administrators.
- Categories include:
1. Application
2. DDoS
3. Normal
4. VPN
5. TOR
6. Snapchat
7. Streaming Services
8. IoT Device Types
9. Anomaly Detection
10. Gaming Traffic
11. Social Media Traffic
NeuralNetIoT: Smart Data 12. Email Services
CPG 78Classification for IoT Networks 13. Cloud Storage
Coconut plantation is a vital agricultural practice in India, economically sustained by sociocultural factors. Despite the existence of various coconut-picking devices,
challenges such as human intervention, lack of user-friendliness, high costs, and unavailability for small-scale operations persist. Therefore, our project aims to develop
CPG 79CocoHunter a device addressing these issues.
This project introduces a sophisticated safety system in mining, primarily focusing on miner protection and operational safety. By integrating ESP modules in mining
equipment, we utilize sensors for real-time monitoring of various environmental factors like temperature, air quality, and potential flooding. This system also includes
vibration and ultrasonic sensors for machinery and debris detection, and a gas sensor for harmful gas detection. Additionally, miners will wear camera-equipped
helmets for immediate hazard identification. While the system includes ore analysis as a secondary feature, its primary goal is to ensure a significantly safer and more
CPG 80Smart Mining Safety System efficient mining environment through advanced technological integration.
● Several challenges, including insecure design and development practices, weak authentication and authorization mechanisms, insufficient encryption, and potential
risks within the customer and data storage, plague the security of IoT devices.

● In this project, we aim to resolve the authentication, data encryption and data filtering problems faced in the IoT security domain and ultimately protect our device
from external attacks.

SECURPI - Security Enhancement, ● We plan to implement a firewall using Raspberry Pi to tackle the filtering problem, secure the network, and use encryption algorithms to provide integrity.
Control, and Unification with
CPG 81Raspberry for Protected IoT ● We also aim to safeguard the IoT device from internal attack and manage access only through cloud and application.
Hemlock: Defending Artists Against
Text-to-Image Models' Style Hemlock protects artists from style replication by generative AI, preserving their creative authenticity in the digital era. It is a machine learning model which
CPG 82Replication manipulates the art such that changes are not visible by human eye, yet effectively prevents generative AI extracting important features from it.
This project leverages smartphone sensors (accelerometer, and GPS) to detect road anomalies like potholes and map their locations in real time. Using advanced
algorithms and machine learning, the system classifies pothole severity and provides actionable insights for road maintenance. A user-friendly mobile app allows civil
Driveaware: For detecting Road authorities to monitor road conditions, plan repairs, and optimize resources. Scalable backend infrastructure ensures efficient data processing and visualization,
CPG 83anomalies promoting safer roads and sustainable infrastructure management.
The scope of this project includes the design, development, and testing of an e-nose system equipped with a sensor array capable of detecting multiple gases. The
system will incorporate machine learning algorithms to enhance its accuracy and adaptability across different environments. Detection of cooking and fire smoke and to
CPG 84Electronic Nose detect food spoilage.
CPG 85FitFruit Developing a real time system for detection of artificially and naturally ripened fruits
 Capstone projeCts 2024

The "Fasal Kavach" capstone project aims to address the challenges faced by the agricultural sector, such as outdated farming practices, lack of knowledge about
modern techniques, and limited access to technology, resulting in significant yield losses. This project integrates IoT technology, deep learning algorithms, and real-
time data analysis to revolutionize traditional farming practices. By utilizing sensors to measure soil pH, humidity, and NPK values, and by using deep learning
algorithms for disease, pest and weed detection, and sending real-time data to farmers through an application, the project empowers farmers to make informed
decisions and take timely actions. The integration of precision cameras for monitoring, and the deployment of AI on NVIDIA's Jetson Nano enhances the system's
efficiency and sustainability. The project aims to create a comprehensive solution that supports farmers in achieving sustainable, profitable, and efficient agriculture,
CPG 86FASAL KAVACH ultimately improving crop yields and reducing the use of pesticides.
Project Description:
This project aims to explore the combined effects of yoga and Sudarshan Kriya (SK) meditation on brain signals to mitigate stress and improve mental well-being. In
today's fast-paced lifestyle, stress has become a prevalent issue, leading to various mental health challenges like depression and anxiety. The study will investigate EEG
signals before and after a three-month regimen of combined yoga and SK meditation practice. EEG signals will be captured from both control and study groups, and
Discrete Wavelet Transform (DWT) will be utilized to decompose the signals into six sub-bands for in-depth analysis. Statistical parameters such as variance, standard
deviation, kurtosis, zero crossing, maximum, and minimum will be computed from each sub-band. These parameters will be subjected to statistical validation using the
Kruskal–Wallis test to ensure their robustness. Subsequently, an Artificial Neural Network (ANN) will be trained using the extracted statistical parameters to classify
subjects as either meditators or non-meditators.

Solution Proposed:

Data Collection: Pre and post-EEG signals will be collected from control and study groups before and after three months of combined yoga and SK meditation practice.
Signal Processing: Discrete Wavelet Transform (DWT) will be employed to decompose EEG signals into six sub-bands for detailed analysis.
Feature Extraction: Statistical parameters including variance, standard deviation, kurtosis, zero crossing, maximum, and minimum will be computed from each sub-
band.
Validation: The obtained parameters will undergo statistical validation using the Kruskal–Wallis test to ensure their reliability and significance.
Classification: An Artificial Neural Network (ANN) will be trained using the extracted statistical parameters to classify subjects as either meditators or non-meditators.
Evaluation: The effectiveness of the classification model in distinguishing between meditators and non-meditators will be assessed, providing insights into the impact of
yoga and SK meditation on brain activity and stress management.
Outcome:
EEG signal based classification The project aims to shed light on the potential benefits of combined yoga and SK meditation practices on mental well-being, as reflected in EEG signals. By
before and after combined Yoga and investigating the alterations in brain activity patterns before and after the meditation regimen, the study will contribute to the scientific understanding of meditation's
CPG 87Sudarshan Kriya effects on stress reduction and mental health enhancement.
Node-ReVault is a decentralized platform for secure and reliable node validation in blockchain systems. It utilizes Proxy Re-Encryption (PRE) to handle data securely,
Node-ReVault Node Validation ensuring its accuracy and reliability during validation. The platform incorporates Proof-of-Stake (PoS), where PRE supports trust and fairness among participants. By
Platform using Proxy Re-Encryption using smart contracts to automate task management, data validation, and reward distribution, Node-ReVault addresses critical challenges like safeguarding data and
CPG 88in Blockchain maintaining its integrity in a decentralized framework.
CPG 89Air Pollution Detection .
Skill Voyage is a cost-effective VR application designed to enhance children's motor skills through interactive, gamified experiences. It uses hand tracking and real-
time interaction within a virtual environment, eliminating the need for expensive proprietary hardware. The integration of NPCs (Non-Player Characters) adds dynamic
CPG 90Skill Voyage interactions and personalized challenges, further enriching the immersive learning experience.
Definition:
A prostate cancer examination website leveraging artificial intelligence to analyze scans submitted by patients which will be later verified by the doctor, assisting
doctors in early detection. The website aims to enhance diagnostic accuracy and streamline the assessment process.

Solution:
The website employs advanced AI algorithms to analyze prostate scans, identifying potential cancerous abnormalities and generating a report. It provides doctors with
real-time insights, highlighting areas of concern and offering suggestions for further investigation. Additionally, it includes a user-friendly interface for patients to
submit scans securely.

Outcome:
The project's outcome is an innovative tool that aids healthcare professionals in early prostate cancer detection, shorten the time of waiting for patients and reducing the
workload of doctors. By harnessing AI capabilities, the website contributes to more accurate diagnoses, facilitates timely interventions, and ultimately enhances the
CPG 91NeuroProsta overall effectiveness of prostate cancer screenings.
Project Definition: Smart Helmet for Miners

The Smart Helmet for Miners is a cutting-edge capstone project focused on elevating mining safety and efficiency. This innovative helmet integrates real-time
monitoring using IoT and sensors to detect environmental hazards, providing instant alerts to miners and control centers. Communication and navigation features
CPG 92Smart Helmet for Miners enhance coordination and emergency response. The outcome anticipates a safer work environment and increased operational efficiency.
Introducing a revolutionary medicine dispensing machine designed for rural areas. Users can select their health issues, initiating a video call with a doctor who provides
CPG 93Mediease a prescription. Once prescribed, users can conveniently collect their medications from the machine, bridging the gap between healthcare access and rural communities.
A rover-based solution that can be used for rescue operations in earthquake, mining, and chemical hazards. While drone-based systems help analyse terrain from an
aerial view, a rover-based system can capitalise on structural damage mapping from a ground perspective. Our proposed solution is based on IoT and ML domains
PRALAYA: UGV-based Disaster using various sensors to analyse the rescue operation before human interference to help form rescue strategies. Not only will it describe the structure damage, but it will
CPG 94Surveillance assist in damage control by tracing the location of victims.
The project involves a glove type utility which would prove to be a better replacement to the ordnary pc mouse, no only perfoming all functions of a mouse but also
capable of many new type of shortcuts simply through gestures.
This can be achieved using parts of Image Processing, Machine learning and IoT
GRIP: Glove-Responsive Interaction Final output of this project is a glove that would have potential to replace mouse completely and provide better user experience. It can have a lot of application in
CPG 95Platform teaching field, where teaching through slides and writing notes would become immensely simplified
Definition:

The SmartDose AI Dispenser is an innovative medical device designed to automate and optimize the process of dispensing medications to patients. It integrates
advanced artificial intelligence (AI) technology to tailor medication administration, improve adherence, and enhance overall patient care.

Solution Proposed:

Automated Medication Dispensing: The SmartDose AI Dispenser automates the dispensing of medications according to prescribed schedules, reducing the risk of
human error and ensuring timely administration.
Connectivity and Integration: The dispenser is designed to integrate with other healthcare systems and devices, facilitating seamless data sharing and collaborative care
management.

Outcome:
The SmartDose AI Dispenser aims to achieve several positive outcomes in healthcare:

Improved Medication Adherence: The automated and personalized approach to medication administration helps patients adhere to their prescribed regimens, leading to
better treatment outcomes.

Enhanced Treatment Efficacy: AI-based dosing optimization ensures that patients receive the right dosage based on their unique health profile, potentially improving
the efficacy of the prescribed medications.

CPG 96SmartDose AI Dispenser Reduced Medication Errors: Automation reduces the likelihood of human errors in dispensing medications, contributing to a safer healthcare environment.
The project aims to develop an integrated system for diabetes management, combining real-time glucose monitoring with a personalized diet plan. This solution
"DiabetesCare+: A Comprehensive involves both hardware and software components to empower diabetic patients in monitoring and controlling their blood glucose levels effectively. The system
System for Glucose Monitoring and provides users with insightful data, personalized dietary recommendations, and support for making informed lifestyle choices. Additionally, it facilitates communication
CPG 97Personalized Dietary Management" between users and healthcare professionals, fostering a collaborative approach to diabetes care.
 Capstone projeCts 2024

Description:
This project aims to develop a machine learning model for predicting the effects of mutations on proteins. By analyzing protein structures and mutation data, the goal is
to provide insights into the functional consequences of genetic variations.

Solution Proposed:
Create a predictive model using machine learning techniques, integrating bioinformatics tools. The model will assess mutation impact, offering a valuable tool for
researchers and clinicians in personalized medicine.

Outcome:
Robust Predictive Model
Predicting the effects and Enhanced Understanding of Protein Mutations
consequences of mutations in Practical Tool for Researchers and Clinicians
CPG 98proteins Contribution to Targeted Therapies and Personalized Medicine.
Problem Definition:
Many individuals with severe physical disabilities face challenges in navigating their surroundings independently, hindering their mobility and overall quality of life.
Traditional wheelchair control interfaces may not be suitable for those with limited motor functions. Consequently, there is a pressing need for an intuitive and efficient
control system that can empower users with diverse mobility restrictions.

Proposed Solution:
We present a Brain-Controlled Wheelchair (BCW) as an innovative solution to address the aforementioned challenges. The system relies on cutting-edge
electroencephalography (EEG) technology to detect and interpret brain signals associated with specific commands. A user-friendly interface, equipped with machine
learning algorithms, translates these signals into precise control commands for the wheelchair's movement.

Expected Outcomes:
1. Enhanced Independence: NeuroNavChair aims to significantly increase the independence of individuals with severe physical disabilities, allowing them to navigate
their surroundings with minimal assistance.
2. Improved Quality of Life: By providing a reliable and intuitive means of mobility, the NeuroNavChair contributes to an improved overall quality of life for users,
fostering a sense of autonomy and freedom.
3. User Adaptability: The machine learning algorithms enable the system to adapt to the unique characteristics of each user, ensuring a personalized and efficient
control experience.

Implementation and Future Developments:

It is designed with scalability in mind, allowing for future upgrades and integration with emerging technologies. Collaborations with healthcare professionals and user
feedback will drive continuous improvement and customization, ensuring the device remains responsive to the evolving needs of its users.

NeuroNavChair: Empowering In summary, the Brain-Controlled Wheelchair represents a groundbreaking solution to empower individuals with severe physical disabilities, offering them a new level
CPG 99Mobility Through Thought" of autonomy and freedom in their daily lives.
The Smart Pantry Manager project aims to transform the management of kitchen pantries with its innovative mobile application, by including features like easy
inventory tracking, alerting users about product expiry dates, and facilitate meal planning by suggesting recipes based on available ingredients.
The proposed solution involves harnessing OCR technology, integrating a sophisticated recommender system, and deploying real-time notifications.
The expected outcome is a significant reduction in food waste, streamlined grocery shopping, and enhanced meal preparation method, making the Smart Pantry
Smart Pantry Manager
CPG 100 Manager an indispensable tool for modern households.
Smart route prediction for Women's 1.Developing a real-time mapping system to guide women through the safest routes during odd hours.
Safety
CPG 101 2 Utilizing data-driven insights to create a dynamic safety map tailored for women traveling alone, offering secure and reliable navigation.
Project Definition:
Dyslexia is a learning disorder that affects reading, writing, and spelling abilities, making traditional learning methods challenging for affected individuals. This
platform seeks to address these challenges by leveraging artificial intelligence and inclusive design principles to create a supportive and accessible learning environment.
Solution Proposed:
A platform - Dokkai Quest, that transforms the challenges of dyslexia into an exciting adventure. With innovative gameplay, interactive activities, and a virtual sensei,
Dokkai Quest revolutionizes the way individuals with dyslexia enhance their reading, writing and learning skills. Some features include Interactive Learning Modules,
Personalised Learning Paths, Speech Recognition and Dictation , all incorporated with GenAI , Computer Vision and NLP.
Outcomes:
Our platforms aims to achieve Enhanced Accessibility by leveraging AI technology and inclusive design principles, improved Learning Outcomes by catering to
individual needs and preferences and Empowerment and Confidence through its supportive features and user-centric design, the platform empowers individuals with
D
CPG 102 okkai - Empowering Dyslexic Minds dyslexia to overcome challenges and build confidence in their academic abilities.
- Problem: Traditional manufacturing faces inefficiencies and downtime due to manual OEE monitoring and lacks compliance with ISO22400 standards.
- Solution: The Intelligent OEE Guardian integrates IoT devices and virtual software for real-time monitoring, remote access, and data analysis, aligning with
ISO22400's OEE indicators.
- Outcomes:
-Real-time Monitoring: Continuous insight into machine performance.
-Remote Access: Stakeholders can access data remotely for timely decision-making.
-Efficient Data Analysis: Advanced analytics tools provide actionable insights.
Intelligent OEE Guardian: A Low- -Customized Dashboards: User-friendly interfaces for various stakeholders including laymen, managers, workers, and CEOs.
Cost IoT-Based Monitoring System -Compliance with ISO22400: Adherence to industry regulations and standardized key performance indicators.
for Manufacturing Industries
CPG 103 - Regulation: Aligns with ISO22400 standards, emphasizing Overall Equipment Effectiveness (OEE) as a unified approach to manufacturing operations management.
Ordinarily image detection works by using Convolutional Neural Networks. Here, several RGB images are used to train the model to be able to predict images. For a
security application however, this is not enough because we need to be able to detect a face in adverse conditions namely in low light conditions and conditions where
the face is obstructed.

In low light, the model will fail to pick up facial features on account of shadows, lack of contrast and lack of color. Similarly if the face is obstructed either
intentionally or unintentionally, the model has only partial view of the face to work with. Both these problems severely limit the ability of a model to predict.

through our project we will implement the following.

1. Image translation: Use Generative Adversarial Network(GANs) or Variational Autoencoders (VAEs) to learn mappings between different image domains. We will
Enhancing Security and Surveillance: train the model to convert images from near-infrared, thermal or partially covered domains to visible spectrum which represents the original face.
Translating Disguised Thermal Facial 2. Facial Occlusion Handling: We will develop algorithms to detect and localize facial occlusions in partially covered images. we will employ inpainting techniques
Images into Visible Faces
CPG 104 such as Context Encoders or Generative Adversarial Networks with an inpainting component, to fill in the missing regions of the face.
The traditional metro uses cards, tokens and scanners for entry and exit. 'Fast Face Metro Transit' eliminates these physical needs by employing a cutting edge facial
Fast Face Metro Transit
CPG 105 recognition technology. It also facilitates automatic calculation and deduction of fare from a wallet making it a completely seamless experience.
In the face of limited visibility caused by smoke during fire incidents, the AgniPath project presents a pioneering solution. Leveraging advanced AI-ML technology, it
AgniPath: Enhanced Rescuer integrates an intelligent de-smoking system designed to enhance real-time video feeds of fire-affected areas. This innovative approach aims to not only improve
Assistance
CPG 106 visibility for rescue operations but also provide critical assistance to rescuers navigating challenging environments.
Exploring Solutions through LLMs for This project explores innovative applications of Large Language Models (LLMs) in solving real-world problems. By leveraging LLMs' advanced natural language
Real-World Contextualised Problems processing, it aims to enhance efficiency and accessibility across sectors, bridging the gap between AI technology and practical solutions for everyday challenges.
CPG 107
SilentSpeak leverages deep learning technology to develop a cutting-edge software solution for seamless communication between deaf individuals and the hearing
community. By harnessing advanced algorithms, it transcends geographical and linguistic barriers, facilitating real-time translation between spoken language and sign
language. SilentSpeak removes obstacles between different countries' sign languages, ensuring accessibility and inclusivity on a global scale. The outcome is a robust
Silent Speak
CPG 108 and intuitive application that empowers users to connect and interact effortlessly, enhancing everyday communication experiences.
This project intends to create an EEG-controlled wheelchair for those with movement disabilities that provides intuitive and direct control via their brainwaves at a
minimal cost. This project also seeks to investigate and implement several DL architectures and signal processing techniques, with the goal of determining which ones
EEG Controlled Device
CPG 109 work best on our project in terms of accuracies and precision.
 Capstone projeCts 2024

Project Definition:
Krishi Seva is a transformative agricultural technology project aimed at addressing the diverse challenges faced by Indian farmers. The project goes beyond traditional
farming approaches by integrating smart technologies, Internet of Things (IoT) sensors, and machine learning to provide farmers with a comprehensive solution.

Solution Proposed:
Krishi Seva introduces a transformative mobile application as a comprehensive solution for Indian farmers, integrating advanced technologies and smart farming
practices. Additionally, the project extends its scope to include hydroponics. Hydroponic farming is currently playing a significant role in revolutionizing the
agricultural industry in India.

Outcome:
The implementation of Krishi Seva envisions a paradigm shift in Indian agriculture. By leveraging technology and innovative practices, the project aims to enhance
Krishi Seva - Empowering Fields, overall productivity through informed decision-making and resource optimization. Farmers using Krishi Seva will benefit from financial empowerment through
Enriching Lives.
CPG 110 predictive analytics, while the integration of hydroponics promotes sustainability and eco-friendly farming practices.
Develop a VR system for exposure therapy to treat specific phobias.
Virtual Reality Therapy for Phobias
CPG 111 Integrate VR hardware with software that simulates scenarios related to the phobia, gradually exposing users to their fears in a controlled environment.
TranspariMed: Advancing Medical The project pioneers the fusion of cutting-edge AI with transparent explanations to revolutionize medical imaging analysis. By integrating Explainable AI (XAI), our
Image Analysis through XAI
CPG 112 system enhances anomaly detection, aiding healthcare professionals in making informed diagnostic decisions, ultimately improving patient care.
We plan to create a robot for hospitals that is capable of :
a) sanitisation of hospital rooms reducing the contact of workers with infectious pathogens,
b) bringing medicines ( and possibly meals ) to bed ridden patients on a regular basis, thus reducing human efforts used up in this simple yet tedious task.
The robot is expected to have the following capabilities:
1) Able to find its path to specific rooms and patients' beds.
2) Able to detect and avoid obstacles
UV-Medi BOT
CPG 113 3) Able to clean/disinfect rooms, and carry and bring medicines to patients.
RescueRover is an innovative autonomous robot developed for search and rescue missions, particularly in post-disaster environments such as collapsed buildings and
other hazardous areas. The robot is equipped with a high-resolution thermal camera, enabling it to detect human presence even in challenging conditions like darkness,
heavy smoke, and extreme weather. In addition to thermal imaging, RescueRover features advanced sensors, including ultrasonic sensors, for effective obstacle
detection and navigation. This capability allows it to autonomously maneuver through cluttered and unpredictable environments, identifying and locating survivors
quickly and safely. The robot’s design is optimized to operate independently, providing real-time feedback to rescue teams and ensuring efficient and effective search
operations without the need for constant human intervention.

RescueRover is also fitted with an onboard display screen that communicates critical information to survivors, such as the nearest extraction points and estimated
arrival times of rescue teams, offering reassurance and guidance. The robot's user interface provides operators with live feeds from the thermal and RGB cameras, GPS
coordinates of detected humans, and data from gas sensors, enhancing situational awareness and enabling timely decision-making. Built with a commitment to
enhancing rescue operations, RescueRover integrates advanced technology with user-centered design, paving the way for more efficient and effective responses in
RescueRover
CPG 114 emergency and disaster scenarios
A Secure Sensor Data Sharing An encryption based healthcare framework will be developed to address the problems related to the sensor-based patient vital body parameters collection, monitoring,
Framework for IoT.
CPG 115 secured data storage and sharing among different stakeholders.
CrisisChain is a cutting-edge disaster management platform designed to
harness the power of blockchain technology and innovative digital tools. By leveraging
distributed ledger technology (DLT), smart contracts, and real-time data analytics,
CrisisChain aims to revolutionize disaster preparedness, response, and recovery efforts by
CrisisChain
CPG 116 enabling efficient coordination, transparent communication, and effective resource allocation.
Smart Walking Stick: An IoT Solution for the Visually Impaired
Project Overview:
SenseWalk: IoT Walking Stick for The aim of this project is to design and develop a smart walking stick equipped with IoT technology to assist visually impaired individuals in navigating their
Safe and Independent Travel
CPG 117 surroundings safely and independently. The walking stick will integrate various sensors, connectivity modules to provide real-time feedback and assistance to the user.
Problem Definition:
Professional athletes constantly seek ways to enhance their performance, and correct technique is crucial for success in various sports.
Proposed Solution:
We propose developing a machine-learning model that analyzes human poses during professional sports performances. The model will be trained on a carefully curated
dataset, focusing on diverse sports scenarios and emphasizing the correctness of the technique. The system will assess the pose correctness and provide insights into
recovery needs in case of injuries.
Expected Outcomes:
XPOSE: Revolutionizing Performance Our project analyzes how people move during training, providing instant feedback on the correctness of their movements. This helps athletes make quick adjustments
with Pose Analysis and Injury during practice.
Insights in Professional Sports
CPG 118 Additionally, our project can assist athletes in recovering after an injury by continuously monitoring their progress through their postures during the recovery phase.

Description:
CollegeReferral revolutionizes the college-to-career transition by combining the power of personalized referrals with cutting-edge machine learning technology. In
today's competitive job market, college students often face challenges in securing internships, projects, and job opportunities that align with their career aspirations.
Conversely, seniors within the same academic community possess valuable insights and experiences that can significantly benefit their juniors' professional growth.
CollegeReferral addresses this gap by providing a platform where seniors can offer personalized referrals to juniors based on the analysis and ranking of student
resumes conducted by machine learning algorithms.

Key Features:

1. Profile Creation and Resume Upload: Students create comprehensive profiles and upload their resumes, highlighting academic achievements, extracurricular
activities, internships, skills, and career aspirations.

2. Machine Learning Resume Ranking:CollegeReferral employs advanced machine learning algorithms to analyze and rank student resumes based on relevant criteria.
Factors considered may include GPA, relevant coursework, project experience, internships, skills, and extracurricular involvement.

3. **Senior Referrals:** Seniors within the same academic institution utilize the machine-generated resume rankings to identify promising candidates among their
juniors. They can then provide personalized referrals, recommending students for internships, projects, job opportunities, or mentorship based on their profiles and
ranking scores.

4. Outcome Tracking: The platform facilitates transparent communication between seniors and juniors by providing a tracking system for referrals. Juniors can monitor
the progress of their applications and engagements resulting from referrals, providing valuable feedback to seniors.

5. Networking Events and Workshops: CollegeReferral organizes virtual networking events, career fairs, and workshops where students can interact with seniors,
alumni, and industry professionals. These events foster meaningful connections and provide additional opportunities for mentorship and career guidance.

Problem Addressed:
College students often struggle to navigate the complexities of the job market and secure opportunities that match their career goals. Similarly, seniors possess valuable
insights and experiences that can significantly benefit their juniors' professional development. However, there is often a disconnect between students and seniors within
CollegeReferral
CPG 119 the same academic community, hindering the exchange of valuable information and opportunities.
Project Definition:
Our project aims to utilize satellite data for forecasting, enhancing short-term prediction accuracy using ThapSat.

Solution Proposed:
We plan to develop a predictive model using machine learning algorithms, integrating advanced satellite imagery and atmospheric data analysis techniques.

Final Outcome:
Forecasting operations using ThapSat Our project will deliver a robust predictive model offering actionable insights for improved decision-making across various sectors reliant on weather information.
CPG 120
 Capstone projeCts 2024

A biometric system relying on a single biometric modality, such as fingerprints alone, often faces challenges due to factors like poor data quality or limited scalability.
Advancing Biometric Recognition: Multibiometric systems, on the other hand, leverage fusion principles to integrate information from various sources, aiming to enhance recognition accuracy and
Unveiling the Power of overcome the constraints of single-biometric systems. Over the past two decades, numerous biometric fusion schemes have been developed. In this biometric fusion
Multibiometric Systems through project, our focus will be on addressing three key questions: what to fuse, when to fuse, and how to fuse for gait, fingerprint, iris, and face, with the aim of designing
Fusion Innovation
CPG 121 robust Multibiometric systems.
Our innovative agricultural technology solution comprises three integral components aimed at revolutionizing farming practices and fostering sustainability. The first
component is the Plant Leaf Disease Detection system, which harnesses advanced image processing and machine learning techniques to swiftly identify and diagnose
plant diseases. By detecting diseases early, farmers can implement targeted interventions, preventing the spread of infections and minimizing crop losses. The second
component is our Pesticide Management Platform, which offers farmers guidance on safe and sustainable pesticide usage. This platform monitors pesticide residues,
BotaniGuard: Defending Plant Life ensuring food safety standards are met while minimizing environmental impact. Finally, our Oxygen-Soil Analysis module optimizes soil health by measuring oxygen
with Advanced Disease and Pesticide levels and providing valuable insights into soil composition. Together, these interconnected modules empower farmers with data-driven tools, enabling them to
Detection"
CPG 122 enhance agricultural practices, mitigate risks, and cultivate sustainable crop production methods for a brighter agricultural future.
Project Name: Smart Safety Helmet for Coal Miners with CO Detection

Problem: Traditional miners' helmets lack advanced safety features, leaving workers vulnerable to exposure to harmful gases like carbon monoxide (CO). This project
aims to address this issue by developing a smart helmet equipped with a CO sensor and data logging capabilities.

Target Audience: Coal miners and mining companies.

Project Goals:

Enhance miner safety by providing real-time CO detection and audible alerts.

Improve emergency response times by notifying miners and supervisors of potential hazards.
Enable long-term exposure analysis through data logging for informed safety protocol development.

Solution Proposed:
Integrated CO sensor: Continuously monitors ambient CO levels in real-time.
Audible alerts: Buzzer triggers at pre-defined CO concentration thresholds, prompting miners to evacuate.
Visual indicators: LED lights provide additional visual cues about CO levels.
Data logging: Records CO readings over time for historical analysis and trend identification.
Durable and rugged design: Built to withstand harsh mining environments.

Technology:
Utilizing existing gas sensor modules for CO detection.
Employing microcontrollers for data processing and triggering alerts.
Leveraging data storage solutions like SD cards or cloud storage.

Outcomes:
Reduced miner exposure to CO: Real-time detection and alarms encourage timely evacuation, minimizing potential health risks.
Improved emergency response: Faster reaction times by miners and supervisors due to immediate alerts.
Enhanced safety protocols: Evidence-based decision making for implementing effective safety measures.
Smart Safety Helmet
CPG 123 Potential cost savings: Reduced healthcare costs associated with CO-related illnesses.
RPMS is one of its kind,Groundbreaking application which serves as one stop solution for all your Research insights, we aim to build an E2E application which scrapes
all the data of paper published from across the departments and provides the same in most interactive and informative way you have ever seen.This include many new
age technologies such as NLP etc. Further all this insights will be shared through the medium of Newsletter. Going ahead this can be scaled into a business model
Research Paper Management System where we standardise the template and would be sell the same on subscription model.
CPG 124
OBJECTIVE:
The key objectives include reducing waiting times for students, ensuring punctuality for
classes, creating income opportunities for e-rickshaw drivers, offering real-time information
on ride availability and designing the app to be scalable for future growth. The success of the
project will be measured based on improvements in these areas and positive user feedback.
EXPECTED OUTCOMES:
● A mobile application to:
➢ Reduce wait time with the help of real time location data and quicker
availability of auto-rickshaws for students.
Ez-Ryde
CPG 125 ➢ Minimize ideal time and maximize operational efficiency for drivers.
Project Definition:
FaceMed Scan is a facial recognition-based solution designed to optimize healthcare processes. It aims to streamline patient identification, access control, and medical
record management, enhancing overall efficiency and security within medical facilities.

Solution Proposed:
The FaceMed Scan system employs advanced facial recognition algorithms to quickly and accurately identify patients. It eliminates the need for manual identification
methods, such as ID cards, by securely linking patients' facial features to their electronic health records. The system also includes features like access control, real-time
record management, and robust security protocols.

Outcome:

Improved Efficiency: Reduces time spent on manual identification processes.

Enhanced Accuracy: Minimizes errors associated with manual data entry.
Increased Security: Adds an extra layer of security to medical facilities and patient data.
Better Patient Experience: Streamlines check-in processes, reducing wait times.
FaceMed Scan
CPG 126 Compliance: Designed to comply with data protection regulations and healthcare standards.
A comprehensive health information platform is needed to provide personalized medication recommendations and facilitate informed decision making. Utilizing the
massive and still growing Cord19 dataset and advanced machine learning techniques, the platform will allow users to input health conditions or medication names,
receive tailored treatments, and access relevant research papers and insights. This innovative solution aims to bridge the gap between medical research and everyday
MedInsight Hub
CPG 127 healthcare decisions, improving health outcomes and empowering individuals to take control of their health journey.
Our project will address IoT data security by simulating cyber attacks on sensor data transmission. To fortify the system, we will employ a machine learning model with
Threat detection on IoT model using anomaly detection algorithms that will learn normal patterns and identify deviations in real-time, enhancing the IoT network's ability to detect and respond promptly to
ML
CPG 128 potential security breaches. This approach will ensure the integrity and reliability of the IoT ecosystem in the face of evolving cyber threats.
 Capstone projeCts 2024

Project Title: Vahan-Chakshu: A Real-Time IoT-Based System for Driver Safety Assistance, Proactive Accident Avoidance, and Post-Crash Care
Project Summary:
Vahan-Chakshu is an advanced driver-assistance system (ADAS) that uses various sensors and machine learning to improve road safety. It can detect driver
drowsiness, alcohol consumption, and accidents, and take appropriate actions like triggering alarms, sending alerts, and pinpointing the vehicle's location. The system
aims to prevent accidents, enhance post-crash care, and assist insurance companies.

Here are the key functionalities of Vahan-Chakshu:

Driver monitoring: Uses sensors and computer vision to track driver's eye closure and detect drowsiness.
Accident detection: Employs sensors and algorithms to identify accidents and trigger alerts.
Post-crash care: Sends alerts to emergency services and provides location information.
Alcohol detection: Monitors driver's alcohol level and sends alerts to contacts if it exceeds the limit.
Vehicle tracking: Continuously monitors the vehicle's location using IoT technology.
Overall, Vahan-Chakshu is a promising ADAS with the potential to significantly improve road safety and accident response.
Output:
Expected Output and Outcome of the proposal:

Deliverable in the form of Hardware: Vahan Chakshu: A low cost IoT based embedded system for real-time monitoring of driver’s behavior and emergency response
system.
Deliverable in the form of Software: 1- Software to detect driver behavior.
2- An efficient algorithm to create Tamper-proof internal and external video footage. an eco-system for secure storage of vehicle data on cloud
Vahan Chakshu
CPG 129 3- Mobile app to assist rescue teams for immediate response and assistance.
In the digital age, organizations grapple with vast amounts of data, including sensitive information stored in
documents. Extracting valuable insights from such documents while maintaining data privacy and security
is paramount. The DataQuest.ai project aims to address this challenge by developing a secure system that
enables query-based tabular data extraction and analysis from various documents and visualizes this
information through interactive dashboards. By combining document analysis and dashboard generation in a
single platform, DataQuest.AI offers a seamless and efficient solution for organizations to make informed
decisions based on their data, while also prioritizing data redaction and incorporating a voice assistant for
query response.
Key Features:
 Document Parsing: Utilize natural language processing (NLP) techniques to parse documents and
identify tabular data.
 Query Interface: Develop an intuitive query interface where users can input their financial queries.
 Data Analysis: Implement dashboard analysis for analyzing extracted data based on user queries.
 Summary Report Generation: Automatically generate summary reports summarizing key financial
insights and recommendations.
 Voice Assistant Integration: Implement a voice assistant feature to enable users to interact with the
system using voice commands and receive spoken responses for query results and summaries.
 Sensitive Information Protection: Develop a feature that conceals company sensitive information like
logo, mails etc from your financial documents, allowing you to share them without revealing sensitive
Document-Derived Financial Insights business details.
with Data Redaction
CPG 130
intuitive web application. The system aims to monitor key parameters such as soil moisture, humidity, and pH levels to assess the plant's health and environmental
conditions. When required, the system will automate the watering process to maintain optimal moisture levels. Additionally, users have the flexibility to manually
control the watering process through a user-friendly web app dashboard.

*Proposed Solution:*

1. *Sensor Integration:*
- Implement soil moisture sensors, humidity sensors, and pH sensors to continuously monitor the plant's environment.
- Ensure sensors are calibrated for accurate measurements, providing real-time data on plant conditions.

2. *Automated Watering System:*

- Develop an automated watering system that activates when the soil moisture level falls below a predefined threshold.
- Integrate a water pump mechanism to deliver water to the plant, stopping when the optimal moisture level is reached.

3. *Manual Control through Web App:*

- Design a responsive web application accessible from various devices, allowing users to monitor and control their plants remotely.
- Provide a manual watering feature, enabling users to initiate watering sessions at their discretion.

4. *Plant Statistics Dashboard:*

- Create a comprehensive dashboard within the web app to display plant statistics.
- Include visual representations of soil moisture, humidity, and pH levels over time, aiding users in understanding the plant's growth conditions.

5. *User Notifications:*
- Implement a notification system to alert users when the plant requires watering or if any environmental conditions are outside the optimal range.
- Notifications can be delivered via the web app or through email to ensure timely responsiveness.

6. *User Authentication and Security:*

- Implement secure user authentication to protect the user's data and ensure privacy.
- Utilize encryption protocols for data transmission between the web app and the plant care system.
Automated Plant Care System
CPG 131
Problem Definition:
The current landscape of virtual conferences poses a significant accessibility challenge for individuals who communicate using Indian Sign Language (ISL). There is a
pressing need for a comprehensive and seamless solution that integrates with popular web browsers (like Chrome), to provide real-time sign language interpretation for
virtual conferences. This extension should not only convert ISL into text but also transcribe spoken language, making the conference content accessible to both the
Deaf and hearing participants. Additionally, to enhance inclusivity, the system must support the display of transcriptions in multiple languages (like Hindi, French and
German ), allowing users to customise the language of the text based on their preferences. The primary objective is to bridge the communication gap in virtual
conferences and create an inclusive environment for all participants, regardless of their preferred mode of communication or linguistic background.

Proposed Solution:
1.To develop a Chrome browser extension for virtual conferences.
2.To address accessibility challenges for individuals using Indian Sign Language (ISL).
3.To provide real-time sign language interpretation within the conference environment.
4.To convert ISL into text and transcribe spoken language for a comprehensive understanding.
5.To ensure accessibility for both Deaf and hearing participants.
6.To support customization of displayed transcriptions in multiple languages (like Hindi, German and French) based on user preferences.

Outcomes:
1. The ISL Chrome extension successfully integrates real-time Indian Sign Language interpretation and spoken language transcription, fostering enhanced accessibility
SignBridge - A Multilingual ISL and for both Deaf and hearing participants in virtual conferences.
Speech Transcription Extension for 2. With support for multilingual transcriptions, the extension breaks down language barriers, promoting inclusivity by allowing users to customise text language based
Virtual Inclusivity
CPG 132 on their preferences.
 Capstone projeCts 2024

Deepfakes, hyper-realistic synthetic media, blur the lines between fact and fiction. These fabricated videos or images can seamlessly manipulate real people into
compromising situations, and pose a significant threat to individual reputation, national security and trust in information.

Problem Statement:
“The proliferation of deepfake technology has introduced a significant challenge to the online landscape. The ability to manipulate images with such ease fosters the
spread of disinformation, erodes trust in visual media, and has the potential to cause reputational harm to individuals and organizations. This project aims to address
this critical issue by developing a user-friendly web application for deepfake image detection, empowering individuals to discern authentic content from synthetic
fabrications.”

Solution:
We propose a deep learning-based deepfake image detection system. This will:
Leverage cutting-edge techniques to analyze images for signs of manipulation.
Be trained on a comprehensive dataset encompassing diverse deepfake creation methods.
Aim for exceptional accuracy in differentiating authentic content from deepfakes.

Objectives:
To achieve high detection accuracy for Deepfakes.
To ensure robustness against diverse Deepfake techniques.
To minimize computational footprint of the Deep Learning model.
Guarding Truth: A Machine Learning To Develop a user-friendly web application for deep fake detection.
based DeepFake Recognition Model
CPG 133
In contemporary manufacturing, the assurance of product quality is paramount to meet consumer expectations and uphold industry standards. However, the
conventional quality control (QC) processes for products, such as chips and chocolates, often face challenges in terms of efficiency and precision. This project
addresses these challenges through a comprehensive exploration of advanced technologies, specifically focusing on the identification and categorization of defective
items. The primary issue at hand is the need for a more sophisticated and automated QC system that can efficiently detect and classify products. Conventional methods
may fall short in providing real-time and accurate results, leading to potential delays, increased wastage, and the delivery of subpar products to consumers. The
integration of the YOLO (You Only Look Once) algorithm, particularly YOLOV5, and deep learning models offers a promising solution to enhance the QC process.
The primary goal is to overcome the limitations of traditional methods by implementing a system that can not only detect products promptly but also classify them
DL-Powered Quality Control for accurately based on predefined criteria. By addressing these challenges, the project aims to revolutionize the manufacturing industry's approach to QC, ensuring the
Packaged
CPG 134 Food efficient production of high-quality products while minimizing wastage and meeting the demands of discerning consumers.
To make our lives more convenient, this device saves valuable time by detecting the queue length at a location. It then displays real-time data about the estimated
waiting time on a mobile application. Moreover, it can predict the future wait time of a given time slot by analyzing
the previously collected data in case we plan on visiting the location in the future.
Lastly, through the virtual queue feature, we can also book a position/slot in the queue so that we can visit when we are notified that our turn has come instead of
spending hours waiting in line.
This project can prove to be extremely helpful in emergencies, specifically in locations such as hospitals and banks. In addition to this, it can be a great asset during
outbreaks of viruses such as COVID so we can take social distancing measures accordingly. Also, disabled individuals would benefit from this since it is very
Queue Analyzer
CPG 135 inconvenient for them to wait in line.
Traditional field based surveying is very time intensive and can be destructive which is not feasible for large scale applications. We will use UAV based 3D structural
Crop Monitoring using UAV
CPG 136 crop information which is more efficient and uses Machine Learning models.
Empowering collaborative finance with innovation, our project aims to stand at the forefront of blockchain based shared wallet solutions. This web application will not
only integrate blockchain for security and machine learning but also features an intelligent AI chatbot to elevate user interaction and support.

Features:
- Blockchain Security: Immutable ledger ensures tamper-resistant transactions.
- Fraud Detection: Proactive fraud detection for enhanced security.
- Intelligent AI Chatbot: Natural language interaction for user support and guidance.
- User-Friendly Interface: React-based responsive frontend for an intuitive experience.
- Real-time Transparency: Stay updated with notifications and real-time tracking.

Blockchain and ML Powered Shared Our project is not just a shared wallet; it's a comprehensive solution that combines security, intelligence, and user-centric features to help users experience collaborative
Wallet
CPG 137 finance with confidence, knowing that their funds are secure, transactions are transparent, and support, just a chat away.
In this project we will develop effective methods for Underwater Image Enhancement utilizing neural networks. These methods are designed to tackle challenges such
as color distortion, reduced clarity, and loss of detail encountered when capturing images beneath the water's surface. By harnessing the power of advanced neural
network algorithms and manual adjustments, our approach aims to restore the natural appearance of underwater scenes, significantly enhancing visibility and overall
AquaVision: Enhancing Underwater image quality. This innovative technique holds promise across a spectrum of fields including marine biology, oceanography, underwater archaeology, recreational
Imagery
CPG 138 diving, filmmaking, and industries requiring precise visualization of underwater environments for research, documentation, and decision-making purposes.
Project Definition :
In today's digital world, keeping communication safe from unwanted access and changes is important. This idea suggests a new way to make digital communication
more secure by combining image watermarking with smart deep learning.

Proposed Solution :
The plan is to use image watermarking, a method already used for protecting copyrights, to hide special marks within digital content. These marks act like secret codes,
helping to check if the content is real and hasn't been messed with. By adding deep learning to this process, the system can automatically adjust how it hides these
marks based on the type of content and how secure it needs to be.

Expected Outcome :
Mixing image watermarking with deep learning is expected to have some big benefits for making digital communication safer. First off, it can catch any sneaky changes
made to the content while it's being sent, making sure it stays true to its original form. Plus, because deep learning can learn and adapt, the system becomes better at
hiding these marks and works well with different kinds of content and security setups.
Securing Digital Communication:
Advancing Protection with Image Basically, this new idea gives a full solution for protecting digital communication from things like unauthorized changes, people trying to get in where they shouldn't,
Watermarking and Learning Models and stealing content. By using image watermarking and deep learning together, companies can make their digital communication much more trustworthy and reliable.
CPG 139
The project will involve the creation of an influencer identification model. With this project idea of ours, we aim to make it easier for emerging companies to identify
undiscovered and potential talent for hiring or collaboration with their company. Our findings will have important results for companies trying to take advantage of
influencers in the esports industry. The approaches we will create can be used to find possible influencers, learn about their communities, and create engagement tactics
that are specifically targeted to significantly enhance the effectiveness of influencer marketing campaigns. We will use community detection, influence scoring using the
PageRank method, and network analysis to present a thorough picture of the YouTube esports scene. In addition to identifying possible influencers, this strategy will
offer information on the composition and behaviour of their groups.During the first stage, we will become acquainted with the YouTube public API and the YouTube
esports scene. After that, relevant influencers and their communities will be the focus of data collection and refining. Then we will create algorithms to anticipate
IDENTIFYING INFLUENCERS IN THE possible influencers and visualize the community surrounding influencers using graph theory.
ESPORTS
CPG 140 INDUSTRY We think that the techniques used for this project can be improved upon and used in other communities and social media platforms
Real-world underwater images suffer from quality degeneration caused by the scattering and absorption of light propagation. The damage of the detailed textures in
Sensitive data hiding using underwater images shows the negative effect of detection and recognition. To recovery the image visibility and sharpness for the above applications, a new image
differential privacy
CPG 141 enhancement method is proposed for extracting the image textures.
 Capstone projeCts 2024

Problem Description: In India, where only 1% of visually impaired individuals are Braille literate, Braille education remains critically inaccessible. It also lacks
interactivity, real-time feedback, and adaptability, posing a significant barrier to literacy and equal opportunity. Our project addresses these challenges by creating a
comprehensive and inclusive learning experience.

Proposed Solution: Braille-Buddy transforms Braille learning by aiming to provide personalized learning paths that cater to individual needs and pace, boosting
learning efficiency and motivation. It revolutionizes Braille learning by combining a Braille display, interactive sensors, and speech synthesis into an adaptive and
engaging ecosystem.

Key Features:
1. Haptic Learning
2.Personalized Lessons
3.Multisensory Learning
4.Adaptive Quizzes
5.Real-time Feedback
6.Track Performance and generate Reports

Expected Outcomes:
1.Revamp Braille Literacy
2.Improving Affordability and Accessibility
Braille-Buddy: Smart Braille Tutor 3.Enhanced Language Skills
System
CPG 142 4.Engaging Learning
Project Description:
The project aims to create a Peer-to-Peer (P2P) energy trading system using blockchain technology. This system will enable direct energy exchange between
prosumers, eliminating the need for intermediaries. By leveraging blockchain's transparency and decentralization, the project seeks to establish a secure and efficient
platform for energy trading.

Proposed Solution:
The proposed solution involves developing a blockchain-based platform equipped with smart contracts to facilitate automated and secure transactions. These smart
contracts will ensure transparency and fairness in energy trading while addressing concerns related to scalability and security inherent in blockchain technology.

Outcome:
Upon implementation, the P2P energy trading system is expected to revolutionize the energy market by providing a decentralized marketplace for energy exchange.
This will empower consumers to choose their energy sources, reduce costs, and promote the adoption of renewable energy. Overall, the project aims to foster a
Peer to peer energy transfer
CPG 143 sustainable energy ecosystem driven by transparent and accessible trading mechanisms.
The project aims to create an intelligent solution for identifying sarcasm in textual data using advanced machine learning and natural language processing techniques.
Sarcasm Detection System
CPG 144 The developed Sarcasm Detection System will contribute to improved sentiment analysis and mitigate misinterpretation in user-generated content.
The proposed surveillance anomaly detection project seeks to revolutionize traditional surveillance methods by harnessing the power of machine learning and deep
learning techniques. In an era where public safety is paramount, the project aims to develop an advanced system capable of automatically identifying unusual events or
behaviors in real-time surveillance footage. By leveraging diverse datasets and state-of-the-art algorithms, the system will extract relevant features from video data,
Realtime Anomaly detection
CPG 145 such as motion patterns and spatial-temporal relationships, to effectively distinguish between normal and anomalous activities.
the user will give an input using speech/text and the software will identify and classify the patterns of brain activity associated with different types of harmful
communication, this could involve using neuroimaging techniques such as electroencephalography (EEG) or functional magnetic resonance imaging (fMRI) to collect
data from individuals while they are exposed to or producing different types of language. The software will also Investigate the relationship between brain activity and
Harmful brain activity detection the effectiveness of interventions for harmful communication. This could involve comparing the neural responses of individuals before and after undergoing
software using neuro Linguistic interventions such as media literacy training or cognitive-behavioral therapy. We have already started the work on our project and hence we are sure that our project
model and NLP
CPG 146 will function in its full flow.

Title-
Revolutionizing Transit: RFID Bus Carnet System

Objective-
•Enhance efficiency: Streamline fare collection processes, reducing boarding times and minimizing delays for passengers.
•Improve convenience: Offer passengers a hassle-free payment experience through contactless RFID cards or devices, eliminating the need for cash or paper tickets.
•Enhance security: Reduce the risk of revenue loss from fare evasion and ensure accurate passenger counts for safety and planning purposes.
•Enable data-driven decision-making: Gather valuable insights on passenger flow patterns, peak hours, and route optimization through real-time tracking and analysis
of RFID data.
•Enhance passenger experience: Provide a modern and seamless transit experience that encourages ridership and satisfaction with the public transportation system.

Methodology-

•Realtime Passenger tracking on the app

•Location based automatic deduction
•Rash Driving Auto Review using accelerometer and head count
•Portal for Users to review balance and recharge
•Free tickets for women and children (satisfying necessary condition)

Experimental Outcome-
The experimental outcome of implementing an RFID bus system include various measurable outcomes and benefits, such as:

Revolutionizing Transit: RFID Bus •Improved Boarding Efficiency: Reduction in boarding times due to faster fare collection processes, leading to increased overall efficiency of bus operations.
Carnet System
CPG 147 •Increased Revenue Collection: Decrease in revenue loss from fare evasion and unauthorized boarding, resulting in higher revenue collection for the transit agency.
MOTO-Makeover is a project to revolutionize car industry by creating 3D model of car by using phone's camera. This will help users to visualize modifications of car
MOTO-MAKEOVER
CPG 148 in real-time.
MapMitra: Navigate Effortlessly
CPG 149 integration.
User Behavior Analysis for Online Design a platform feature that employs natural language processing techniques to assess user behavior in public chat forums of a streaming service, empowering
streaming Services
CPG 150 streamers with data to manage and customize their viewer base effectively.
Army Men go through a lot of daily struggles includes warfare's, irregular geotracking and a lot more. Our capstone project focuses on enhancing military safety
Vajra: A smart wearable for army through innovative technology. We aim to create a landmine prototype that distinguishes between soldiers and intruders, taking appropriate action. We are planning to
personnel
CPG 151 design a smart wearable for them which will make a whole lot easier for them including rescue time operations and their day to day activities.
 Capstone projeCts 2024

In the traditional notarization process, the reliance on physical stamp paper and in-person presence poses significant limitations in terms of accessibility, efficiency, and
security. The need for physical presence at a notary office and the dependence on paper-based documentation result in cumbersome procedures, delays, and potential
risks of document tampering or fraud. Furthermore, the centralized nature of traditional notarization systems raises concerns about data security and trustworthiness.
The objective of this project is to address these challenges by transforming the traditional stamp paper notarization process into a digital e-notarization system
leveraging blockchain technology. The proposed system aims to eliminate the need for physical presence by enabling users to notarize documents online. This transition
to digital e-notarization will involve the creation of an online repository for document storage and verification, facilitated by blockchain's decentralized and immutable
nature.

2.Existing Approach

Existing Approach for Notarization:

The traditional approach to notarization involves several manual steps and relies heavily on physical documentation and in-person interactions. Here's an overview of
the existing process:
•Document Preparation: The individual or entity seeking notarization prepares the document requiring notarization. This document could be a contract, agreement,
deed, affidavit, or any other legal instrument.
•Visit to Notary Office: The document holder must visit a physical notary office or a designated notary public to have the document notarized. This requires scheduling
an appointment and traveling to the location.
•Identity Verification: At the notary office, the document holder presents valid identification to prove their identity. The notary public verifies the identity of the signer to
ensure they are authorized to sign the document.
•Signing of Document: The document holder signs the document in the presence of the notary public. The notary witnesses the signing process and confirms the
authenticity of the signature.
•Notary Seal and Signature: Once the document is signed, the notary public affixes their official seal or stamp onto the document and signs it. This seal serves as a
certification of the notary's authority and the authenticity of the document.
•Documentation and Record-Keeping: The notary public maintains a record of the notarized document, including details such as the date, time, location, and parties
involved. The document may also be copied or scanned for archival purposes.
•Payment of Fees: The document holder pays the notary fees for the services rendered, which typically vary depending on the jurisdiction and the type of document.
•Return of Notarized Document: After notarization is complete, the document holder receives the notarized document along with any copies or certificates as required.
Sakshya
CPG 152
Our project is aimed at overcoming language barriers through innovative technology. By combining bone conduction technology with language translation algorithms,
our goal is to create a device that facilitates seamless communication between individuals speaking different languages, all this in a day to day wearable Spectacles.

Real Time Language Translating Bone conduction is a technology that transmits sound vibrations to Cochlea of your ear via the bones around your ear. It is not only a safer hearing alternative but also
Spectacles
CPG 153 enables mildly deaf people to hear sounds.
The objective of this study is to enhance the efficiency of Diabetic Retinopathy (DR) detection, a leading cause of vision impairment, by addressing the challenges
posed by manual detection methods. We propose a novel approach utilizing a Convolutional Neural Network (CNN) model. This method aims to overcome limitations
Diabetic Retinopathy Detection in existing automatic techniques, particularly in capturing the nuances of early-stage diabetic retinopathy. This innovative solution holds the potential to significantly
using Neural Network
CPG 154 improve early detection, minimizing the risks associated with delayed diagnosis in diabetic retinopathy.
Project Description:
The project involves creating an app and hardware device that blend positive psychology and mindfulness principles. The hardware device acts as a temporary relief
during panic and anxiety attacks and provides companionship. It includes a panic button feature for audio-guided breathing exercises and SOS alerts. The goal is to
encourage healthy emotional expression and mindfulness.

Solution Proposed:
The solution integrates an app with resources like guided meditations and mood tracking, along with a hardware device for immediate support during distress. By
combining technology and psychological techniques, it aims to empower users to manage emotions effectively and cultivate mindfulness in their daily routines.

Outcome:
The project aims to promote emotional awareness, resilience, and mindfulness, reducing stigma around mental health. Through tools for managing panic and anxiety
Kintsugi - Mental Wellness
CPG 155 and encouraging open expression of feelings, it seeks to contribute to users' overall well-being and help them lead healthier lives.

The project's goal is to create a voice-and gesture-controlled wheelchair system that will increase mobility and freedom for those with impairments. Conventional
wheelchair control systems frequently rely on users entering commands manually using joysticks or buttons, which can be difficult for those with poor hand dexterity or
Vak Chakra
CPG 156 mobility. The goal of this project is to make a wheelchair user-friendly control interface by incorporating gesture and speech recognition technology.

This capstone project aims to thoroughly analyze and critically evaluate the influence of IR components on Retrieval-Augmented Generation (RAG) systems.
Information Retrieval (IR) is a field within computer science that focuses on retrieving relevant information from large collections of data. It involves techniques and
algorithms to search, filter, and rank documents based on their relevance to a user’s query.
The project will focus on determining the characteristics a retriever should possess for effective prompt formulation within RAG systems. The project will analyze the
relevance of retrieved documents about the prompt, considering their position and contextual inclusion. By conducting this analysis, the project seeks to provide
insights into the optimal strategies for integrating retrieval with language generation models, thereby contributing to advancements in this field.
RETRIEVAL AUGMENTED Retrieval Augmented Generation (RAG) systems significantly advance over traditional Large Language Models (LLMs). However, there is a gap in understanding the
GENERATIONS IN LARGE LANGUAGE influence of Information Retrieval (IR) components on the effectiveness of RAG systems. While existing research primarily focuses on the generative aspect of RAG
MODELS
CPG 157 systems, there is limited analysis of the characteristics of the retriever and the type of documents that should be retrieved to enhance RAG performance.
Millions of farmers in India grapple with accessing and understanding critical agricultural information due to language barriers. Traditional information sources often
lack inclusivity in local languages, hindering knowledge dissemination and effective decision-making.

We propose a transformative project that leverages the power of AI to bridge this gap. We will develop an Indic language-based agricultural information system
powered by cutting-edge natural language processing (NLP) techniques as Transformer based encoder-decoder model to understand and respond to farmers' queries in
their native Indic languages (e.g., Punjabi, Hindi, Marathi), empowering them to access vital knowledge through voice or text input.
We can also leverage pre-trained language models (PLMs) like Transformer-based models and fine-tune them on a vast corpus of agricultural text data in relevant
Indic languages.
We will develop a user-friendly interface that seamlessly supports multiple Indic languages for input and output both in textual and voice forms.

Our project holds immense potential to revolutionise agricultural information access for Indian farmers. Collaboration with relevant organisations like agricultural
cooperatives or NGOs to ensure wider adoption and impact among the target audience. Imagine farmers in Punjab being able to ask questions about crop diseases in
Agriculture chatbot for Indic their language of choice and receive accurate answers in their preferred form (speech or text).
Language
CPG 158 By leveraging the power of AI and Indic languages, we can create a more informed, empowered, and productive agricultural sector in India.
TRAFFIQ
CPG 159
 Capstone projeCts 2024

In the past few years, the investigation of novel deep-learning methods has dramatically improved the field of EEG-based motor imagery classification for BCI systems
which is highly essential for such interfaces. In our proposed approach, we aim at improving motor imagery classification from EEG signals with GNN, finding a
manifold in the data dimension with BI-LSTM and constructing a Deep Belief Attention Network (DBAN). These three effective techniques are brought together with
the aim of using in-depth spatial and temporal information as well as increasing dimensionality while at the same time capturing complicated patterns of the EEG data
for better classification performance.
Primarily, GNNs, which are exploited to establish spatial connections between EEG data, take into account the graph-based representation of the relationship between
the electrode. The key advantage of GNNs is carefully modeling the interactions of EEG channels so that the features
crucial for an accurate prediction are extracted. Second, we employ Bi-Directional Long ShortTerm Memory (BI-LSTM) model which is used for multivariable
learning helping to model the temporal features of the EEG signals. BiLSTM models are suitable for those situations
when it is necessary to capture long-range dependency and subtle temporal patterns of EEG time series. Moreover, DBAN, a novel Deep Belief Attention Network
combining the strengths of deep belief networks and attention mechanisms, is also defined. In hierarchical
representations of vital EEG features, the student is taught to pay attention to only the regions of the input data that are of importance. It is attention mechanism that
powers the neural network to detect discriminative actions and consequently boosts classification accuracy. The proposed set-up is evaluated by scouring open EEG
motor imagery data and it is proven to be more effective than existing methods. Given the presented experimental results, it can be stated that the proposed method is
Graph-Based Manifold Learning for highly effective in the accurate classification of motor imagery tasks, thus the issue of using spatial, temporal, and hierarchical representations in EEG-based BCI
Motor Imagery Classification with systems becomes very urgent. The proposed framework holds promise for advancing the development of robust and efficient EEG-based motor imagery classification
Deep Belief Attention Networks
CPG 160 systems with potential applications in neurorehabilitation, assistive technology, and brain-controlled interfaces.
PROJECT DESCRIPTION:-
A pill dispenser project involves the development and implementation of a device or system designed to assist individuals in managing their medication schedule. The
primary goal is to create a reliable and user-friendly solution that helps users organize and take their prescribed medications in a timely and accurate manner.
SOLUTION:-
A smart pill dispenser with automated detection and dispensing capabilities, integrated alarms for user reminders, and caregiver notifications ensures accurate
Pill Dispenser
CPG 161 medication adherence, while its user-friendly design simplifies the process of loading, managing, and taking medicines.
Imagine a shopping experience where every interaction is effortless, every purchase is streamlined, and every moment is focused on what matters most—your
satisfaction. Introducing our innovative solution: the Smart Shopping Cart.
In today's fast-paced world, time is of the essence. With traditional checkout processes often causing delays and frustration, we saw an opportunity to revolutionize the
way people shop. Our Smart Shopping Cart combines cutting-edge technology with unparalleled convenience, allowing customers to scan and pay for their items as
CartEase: Smart Shopping Solution
CPG 162 they shop, eliminating the need for traditional checkout lines.
Project Definition:
In the realm of object detection and criminal studies, efficient image capture from surveillance devices is imperative for accurate identification. However, many
instances involve incomplete or occluded images, making it challenging to discern critical details. The project aims to address this issue by employing Generative
Adversarial Networks (GANs) to perform occluded object detection. Specifically, the initial focus is on face recognition, with the potential for expansion to other
objects in the future.

Solution Proposed:
Generative Adversarial Networks (GANs) will play a pivotal role in this project. GANs consist of two neural networks, a generator and a discriminator, trained in
tandem. The generator creates realistic images, while the discriminator evaluates their authenticity. This adversarial process results in the generation of high-quality,
realistic images, even from partial or occluded inputs.

Additionally, the incorporation of Style GANs enhances the project's capabilities. Style GANs focus on generating images with specific styles, ensuring that the
reconstructed images maintain the original facial features while addressing occlusions. This methodology significantly improves the accuracy and reliability of the
object detection process.

Outcome:
The primary outcome of this project is the development of an advanced occluded object detection system, initially specialized in face recognition. By leveraging GANs
and Style GANs, the project aims to transform incomplete or occluded images into comprehensive, high-resolution representations. This technology will have
significant applications in criminal studies, where often only partial images are available from surveillance footage.

The creation of a dedicated dataset is a crucial aspect of this project, facilitating the training of the neural networks with diverse occluded scenarios. As the project
progresses, the dataset can be expanded to include various objects beyond faces, broadening the scope of applicability.

Occulted Object Detection Using In summary, the Occulated Object Detection Using GANs project strives to revolutionize image recognition in incomplete or occluded scenarios, with a focus on
Generative Model
CPG 163 enhancing criminal studies through the reconstruction of partial images captured by surveillance systems.
1.An Integrated IoT Patient Monitoring System is designed to enhance
patient care by automating the monitoring of two critical aspects of patient
management: intravenous (IV) fluid administration and urine output. The system
incorporates a Drip Bottle Monitoring module and a Urine Bag Monitoring module.

The Drip Bottle Monitoring system enables real-time tracking of IV fluid levels,
generating an alert and a beeping signal when the fluid reaches a predefined threshold, prompting the nurse to stop the regulator. Similarly, the Urine Bag Monitoring
system continuously monitors urine levels, and when the bag approaches its capacity, it triggers a warning signal and an audible beep, alerting the nurse to discard the
filled bag. This dual-system approach aims to minimize the risks associated with IV therapy and urine management, ensuring timely intervention and improving patient
safety.
With real time updates with a camera monitor, it makes it easier to monitor multiple patients at once.
Along with this there will be a full fledged website with database which will be an all in one place for all medical records of the patient and all daily check-in vitals.

MediSync:
An Integrated IoT Patient Monitoring System is designed to enhance
patient care by automating the monitoring of two critical aspects of patient
management: intravenous (IV) fluid administration and urine output. The system
incorporates a Drip Bottle Monitoring module and a Urine Bag Monitoring module.

The Drip Bottle Monitoring system enables real-time tracking of IV fluid levels,
generating an alert and a beeping signal when the fluid reaches a predefined threshold, prompting the nurse to stop the regulator. Similarly, the Urine Bag Monitoring
system continuously monitors urine levels, and when the bag approaches its capacity, it triggers a warning signal and an audible beep, alerting the nurse to discard the
filled bag. This dual-system approach aims to minimize the risks associated with IV therapy and urine management, ensuring timely intervention and improving patient
safety.
MediSync ICU
CPG 164 With real time updates with a camera monitor, it makes it easier to monitor multiple patients at once.
The "Strike-Through Text Identification from Handwritten Text" project aims to develop a system capable of detecting and extracting text that has been crossed out or
struck through in handwritten documents. This involves utilising deep learning techniques to recognise and isolate portions of text that have been marked for deletion
Strike-Through Text Identification or modification. The project could find applications in digitising historical documents, enhancing optical character recognition ocr accuracy, and improving document
from Handwritten Text
CPG 165 analysis in various fields.
SDNUM: SDN-Enabled The project develops an Internet of Underwater Things (IoUT) which explores the suspense and monitors the ocean or underwater region autonomously. The IoUT is
Communication Network for the combination of the Underwater Wireless Sensor Network (UWSN) with the Internet. The IoUT can help track and monitor climate, natural turbulence, pollution
Underwater Monitoring
CPG 166 control, oil industry, military surveillance, etc.
We plan to create an interface that helps the visually impaired perform their daily tasks with ease.
For the same, we intend to design an Android application to identify and locate objects using image recognition in real-time. The object's location shall then be
Object Recognition interface for communicated to the user through audio output.
visually impaired
CPG 167 This shall enable better navigation for those with visual impairment.
A centralized web application for managing of various permissions approval and logistics process of all the societies of our institute. The current system of seeking
permissions involves a hierarchy of approvals to be granted by different stakeholders at each level. The manual process is time consuming and inefficient for both
Society Permissions Portal
CPG 168 students and admins. Our solution aims at modernizing the process, improving efficiency and at the same time promoting sustainability.
 Capstone projeCts 2024

The Quick Buy System aims to transform the shopping experience by integrating advanced technologies such as RFID (Radio Frequency Identification), QR codes, and
a custom Android application. Each product in the store is tagged with an RFID label, while each shopping basket features a unique QR code. Upon entering the store,
customers scan the basket’s QR code using a dedicated Android application, which links the basket to their profile and sets the stage for a fully automated shopping
journey.

As customers move through the store and add items to their basket, the RFID reader embedded within the basket automatically detects and records the RFID tags on
the products. This real-time tracking ensures that the customer's virtual cart in the application is always up-to-date with their selections, eliminating the need for manual
scanning at the checkout counter. This significantly reduces wait times and enhances the overall shopping experience.

The Android application, developed using Java, serves as the primary interface for managing the shopping journey. It allows users to log in, scan QR codes, track the
items in their basket, and complete the checkout process. The backend of the application is powered by a Flask API, which facilitates seamless communication between
the application and the database. This ensures that user data and cart information are synchronized in real-time, providing a responsive and reliable shopping
experience.

Firebase is utilized for real-time database management, ensuring that any changes in the user’s cart are instantly reflected in the application. This real-time
synchronization is crucial for maintaining accuracy and reliability, allowing users to trust that their cart is always up-to-date. Additionally, NodeMCU microcontrollers
manage the RFID readers, further enhancing the system’s reliability by ensuring accurate data capture and transmission to the backend.

A standout feature of the Quick Buy System is its approach to handling age-restricted products, such as tobacco. The system includes a secure locker for these
products, which is equipped with a unique QR code. When a user wishes to access the locker, they scan the QR code using the Android application. The application
then verifies the user’s age based on their profile information. If the user is 18 years or older, the locker is unlocked, allowing access to the tobacco products. This
feature ensures compliance with legal age restrictions and adds an extra layer of security to the system.

The Quick Buy System integrates several key technologies to deliver a seamless shopping experience. Java is used for developing the Android application, offering a
robust platform for user interaction and QR code scanning. The Flask API serves as a lightweight and efficient backend, handling user requests and managing database
interactions. Firebase’s real-time database capabilities ensure that user data is synchronized instantly, while NodeMCU microcontrollers manage the RFID readers,
ensuring accurate and reliable data capture.

Beyond improving the shopping experience, the Quick Buy System has the potential to significantly reduce operational costs for retailers by minimizing the need for
QuickBuy
CPG 169 manual labor at the checkout counter. This allows staff to be reallocated to other areas of the store where they can provide more value to customers. Furthermore, the
AgroVision is an innovative solution designed to help farmers
optimize productivity, resource use, and environmental impact in
agriculture. It combines advanced image processing techniques with
hardware such as sensors, GPS modules, and adjustable cameras to
offer real-time insights into crop health, soil conditions, and yield.
The system includes a semi-automated cart for data collection and an
CPG 170AgroVision automatic pesticide and fertilizer sprayer.
Project Definition- Automated Power Consumption Analysis will transform the way you monitor electricity. The project uses image processing to extract meter data,
which is then fed into a machine learning model for predictive analysis. The findings are presented on an interactive internet dashboard, which includes insights into
usage trends, real-time projections, and comparison data.

Solution Proposed- Use image processing to acquire meter data, then integrate a machine learning model for analysis. The processed data is displayed on a website
dashboard, offering customers real-time predictions, usage trends, and relevant comparisons.

Automated power consumption Outcome- The project's goal is to offer an efficient system for automated power usage analysis, providing users with information for proactive energy management.
analysis
CPG 171 The internet dashboard displays consumption trends, promoting informed decision-making and contributing to sustainable behaviors.
The project aims to develop an interactive kiosk system for different domains which will provide users with information in their preferred language. The kiosk will
utilize speech recognition technology to understand user queries, identify the language spoken, and then retrieve relevant information from a database. A large language
SpeakEasy: A Multilingual model (LLM) will process the query and generate responses in the user's language, ensuring clear and accurate communication. The system will enhance the user
informational kiosk
CPG 172 experience by offering seamless access to essential information.
Project Title:- Car detection in Fog

Project Description:-We will use multiple sensors and lider camera to detect a car travelling in dense fog which will help to avoid accident and driver can drive safely .
We will use Lider camera:LiDAR is an acronym for Light Detection and Ranging. In LiDAR, laser light is sent from a source (transmitter) and reflected from objects in
the scene. The reflected light is detected by the system receiver and the time of flight (TOF) is used to develop a distance map of the objects in the scene. A LCD
screen will also be present to tell the distance pf the travelling Car.Lasers and LED light sources are commonly used in these applications .Infrared light is used detect
the fog safely not only at night but when there is sunlight.Fog sensors work like A detector behind a lens in another window is sensitive to incoming laser light in a
narrow lobe that overlaps the transmitter beam. If there are fog particles in the overlap zone, light is scattered back and a signal is induced on the sensor's raw signal
output.An Ultrasonic sensor will also be present which will be attached to a servo motor which will keep revolving the sensor . There will be a projection screen act as
Car Detection in Fog
CPG 173 a radar system which will detect about the car and will highlight that part so as to know about the correct side.
CartiFied transforms the traditional retail shopping system integrating advanced technologies to create a seamless and automated shopping journey for customers. The
Project aims on applying Computer Vision technology with Human motion detection coupled with sensors and Raspberry Pi. This system tracks the items added and
CartiFied: Shopping Made Smart
CPG 174 removed from the cart updating your app making the bill payment process easier.
Project Definition:
Design and develop AR glasses for military use, focusing on target detection, distance estimation, and friend or foe detection to enhance soldiers' situational awareness
and minimize friendly fire incidents.

Solution Proposed:
Software: Computer vision algorithms for target detection, depth sensing for distance estimation, friend or foe detection algorithms, GIS integration, real-time
processing, and intuitive user interfaces.
Outcome:
Successful deployment of AR glasses empowering soldiers with:

Enhanced target detection and identification.

Improved distance estimation.
Effective friend or foe identification.
Enhanced decision-making and reduced friendly fire incidents.
AR glasses for military use
CPG 175 Positive soldier feedback on usability and effectiveness in combat scenarios.
To execute a comprehensive project titled "Urban Navigator," aimed at predicting metro stops in new cities based on a variety of parameters.Through this project we
plan to develop multiple datasets for different cities, encompassing essential information of specific locations and incorporating a target column indicating the presence
or absence of a metro stop in the city, providing valuable labels for the predictive modelling phase.
Thus leveraging advanced machine learning techniques to analyse the relationships between geographical features and the likelihood of a metro station, resulting in a
Urban Metro Rail Prediction
CPG 176 comprehensive visualisation of potential metro stops.
SPAWC: Solar-Powered Aquatic Solar-Powered Aquatic Waste Collector is an innovative solution designed to combat the ever-growing problem of debris in our waterways. This project leverages
Waste Collector Capstone Project advanced robotics and AI technology to develop an autonomous system capable of efficiently collecting and removing debris, particularly focusing on items like
Proposal
CPG 177 plastics and organic waste from the surface of rivers, lakes and ponds.
 Capstone projeCts 2024

In the medical field, there is a need for efficient and accurate diagnosis and explanation of
medical conditions to patients. This involves interpreting both textual descriptions of
symptoms and medical images, which can be complex and varied across different
languages and medical specialties. Traditional methods of diagnosis and explanation may be
time-consuming, error-prone, and may not cater to multilingual populations adequately.
**Proposed Solution:**
To address these challenges, we propose a solution leveraging Language Models (LLMs),
multimodal learning, and a multilingual approach within the field of medical diagnosis and
explanation (MDE). The system will take in input from users in the form of text and medical
images and provide a summarized description of the diagnosis and its causes.
1. **Multimodal Learning:** The system will be designed to process both textual descriptions
and medical images simultaneously. This involves training models capable of understanding
and synthesizing information from different modalities.
2. **Language Models:** Advanced language models, such as GPT (Generative Pre-trained
Transformer) or its variants, will be utilized to understand the textual descriptions provided by
users. These models can comprehend medical terminology and context, enabling accurate
interpretation.
3. **Multilingual Approach:** Given the diverse linguistic backgrounds of patients and
healthcare professionals worldwide, the system will support multiple languages. This
involves training models on multilingual datasets and employing techniques for language
translation and understanding.
**Expected Outcomes:**
1. **Accurate Diagnosis:** The system aims to provide accurate diagnoses based on the
user's input, integrating both textual and visual information.
2. **Comprehensive Explanations:** Users will receive summarized descriptions of their
diagnosis and its causes, tailored to their understanding level and language preferences.
3. **Efficiency:** By automating parts of the diagnostic process, the system can potentially
Multimodal Multilingual Medical reduce the time required for medical consultations and improve overall efficiency.
Diagnosis and Explanation (M3DE): 4. **Accessibility:** Multilingual support ensures that the system can cater to diverse patient
Enhancing Healthcare Through AI
CPG 178 populations, improving accessibility to healthcare services.
Machine Learning based FoG By integrating multiple detection models and features, the system will provide real-time feedback through a vibration motor to help users manage FoG episodes
detection
CPG 179 effectively.
Project definition: The goal of this project is to develop a hardware machine that can semi-automate the process of
sorting clothes using machine learning. The machine will be controlled by a
microprocessor and will use a combination of hardware design, software programming, and
machine learning techniques to segregate clothes based on their quality.
Outcome: Our working prototype of the model is capable of automated segregation of clothes on the basis of quality.
The segregation process will be faster and more accurate than manual sorting, resulting in time
savings and increased productivity.
Solution: We will be developing a prototype that can conduct continuous monitoring in the conveyor belt
of the clothes production line on the basis of factors defined by us that will define a cloth's quality. The camera tracks clothes on the conveyor belt and
Automated Cloth quality segregation our onboard pc raspberry pi makes the robotic arm decisions. This approach is based on learning
system
CPG 180 from past experiences which had separate systems for inspecting defects and were less efficient.
Our Project is to select a mentor for a capstone Project by building a user-to-client portal. This portal helps the student find the mentor. This portal helps to save time
Mentor assign for Capstone
CPG 181 for the student as well as the mentor.
The project introduces an interactive platform leveraging augmented reality (AR) for learning sign language. By seamlessly integrating AR technology, users can
engage in immersive sign language lessons, receiving real-time feedback on their signing accuracy. The platform offers a user-friendly interface with a variety of
AR-Enhanced Sign Language lessons and exercises, allowing individuals to improve their signing skills at their own pace. The outcome is an accessible and inclusive learning experience,
Learning Platform
CPG 182 empowering users to communicate effectively through sign language.
Project Definition:
The water surface cleaner prototype aims to address the pressing issue of water pollution by developing a cost-effective and efficient solution for cleaning debris from
the water surface. The project seeks to leverage robotics technology to autonomously navigate water environments and collect floating debris while minimizing harm to
marine life and ecosystems.

Solution Proposed:
The proposed solution is to design and build a prototype of a water surface cleaner, a small-scale autonomous robot capable of identifying and collecting various types
of debris from the water surface. It will utilize object classification techniques to navigate its environment and efficiently clean targeted areas. The prototype will
prioritize simplicity, affordability, and environmental sustainability, making it accessible to a wide range of users interested in water conservation efforts.

Outcome:
The outcome of the project will be a functional prototype of a water surface cleaner capable of autonomously navigating water surfaces, detecting and classifying
debris, and collecting it using appropriate collection mechanisms. The prototype will demonstrate the feasibility of using robotics technology for water surface cleaning
and provide valuable insights into the design, development, and deployment of future iterations. Additionally, the project will contribute to raising awareness about the
Water Surface Cleaner
CPG 183 importance of water conservation and the role that innovative technologies can play in addressing environmental challenges.
The project aims to develop a Smart Helmet and Alert System utilizing hardware components such as Raspberry Pi or Arduino. The system is designed to provide real-
time alerts to emergency contacts and local authorities in the event of an accident. The primary goal is to automate the existing accident recovery system, enhancing
Smart Helmet and Alert System
CPG 184 efficiency and response times for emergency situations.
PB-GPT
CPG 185 Leveraging the power of LLMs for regional language Punjabi for various applications to build PunjabGPT (PB-GPT)
Cybersecurity
CPG 186 Cybersecurity
In remote border areas, the existing security measures are often inadequate in effectively detecting and preventing illegal crossings or terrorist activities. Traditional
surveillance systems face challenges such as connectivity issues, limited power sources, and high operational costs. To address these challenges, our project aims to
Beyond Boundaries: LoRaWAN develop a border security solution using LoRaWAN technology integrated with a camera system. This solution will leverage AI/ML algorithms for real-time detection
Integrated Intelligent Border and classification of abnormal activities, such as illegal border crossings or potential terrorist threats. The key focus is to make the alert system as well as the
Monitoring System
CPG 187 communication possible from our system to the control center.
The Smart Infant Monitoring System will be designed to address the challenges faced by many families in baby care. It aims to simplify the process of monitoring
babies by providing high-quality features like Sleep Detection , Cry detection , Temperature and Humidity Measurement for infant's comfort and Infant's expression
detection so that parent or a doctor can understand what the infant is trying to convey.
Smart Infant Monitoring System
CPG 188 Outcome - A hardware model trained on machine learning and deep learning datasets to achieve the stated features.
1. Facilitate Communication: To bridge the gap between individuals who cannot speak and those unfamiliar with sign language by converting hand gestures into speech
in real-time.
2. Promote Inclusivity: To create an accessible platform that empowers individuals with speech and hearing impairments to communicate effectively in diverse social
and professional settings.
3. Enhance Learning: To provide an interactive and immersive 3D learning EdTech platform where users can purchase courses designed to teach sign language using
Blender animations and real-time demonstrations with the glove.
4. Leverage Technology: To integrate advanced technologies like Blender modeling, React, Express.js, Three.js and Arduino-based hardware for delivering a seamless
Development of Indian Sign to and engaging user experience on the platform.
Speech Conversion Glove using Deep 5. Ensure Affordability: To design and manufacture the product at an affordable price while offering reasonably priced courses, ensuring accessibility to a wider
Learning combined with an Edtech audience, including underserved communities.
Platform
CPG 189
 Capstone projeCts 2024

"3D Moments from Near-Duplicate Photos" is a groundbreaking computational photography project that transforms pairs of near-duplicate photos into immersive 3D
videos. By leveraging advanced algorithms, our solution interpolates scene motion between the photos while generating realistic camera motion with parallax,
providing a heightened sense of depth.
Our approach involves representing the scene using feature-based layered depth images augmented with scene flow, enabling precise motion interpolation and
independent control of the camera viewpoint. The outcome is a series of photo-realistic space-time videos with motion parallax and dynamic scene elements, even
accurately recovering occluded regions from the original views.
3D Moments from Near-Duplicate Through extensive experimentation, our system has demonstrated superior performance compared to baseline methods on public datasets and real-world photo
Photos
CPG 190 collections, showcasing its effectiveness in producing captivating 3D moments from ordinary photos.
districts based on the regularly measured water gauge level. Flood warnings are issued by CWC to the state administration, whenever the gauge levels cross the danger
mark. In all such cases, the severity of the problem is essentially indicated in terms of flood levels. However, the major concern of the state administration is regarding
the difficulty faced in assessing the spatial extent of flooding for mapping rescue and relief works. Large volumes of imagery corresponding to varied flood inundations
are expected to be available with NRSA and other freely assessable sources. The past imageries corresponding to a specific forecast site can be collected and each
imagery be stamped with the observed water level of that particular date and time. An AI-ML-based GIS application can be developed to process the available
imageries for projecting new imageries using suitable interpolation/extrapolation of 2D information corresponding to specific flood levels within a given range. We are
Projection of the extent of expected to develop the above-mentioned application using open-source software and demonstrate its utility for any one of the CWC’s forecast sites (List will be
inundation corresponding to the shared) with readily available imageries. We can also seek to add value by incorporating tools for making critical infrastructure, filtering outliers, pointing out accuracy
forecasts of flood levels in a river.
CPG 191 levels of projected inundation, etc.
The healthcare industry is producing a large amount of data which is highly sensitive. We are required to analyze and understand the associations and patterns in the
data for the drug, vaccine, and critical symptom analysis. Sharing medical data with any third party may cause serious privacy threats. Many privacy-preserving
sensitive patterns hiding techniques exist in the past that mask sensitive patterns before sharing data with others. These approaches delete some part of the
information and ensure that no confidential information gets exposed while sharing or analysing data. However, deleting information causes utility reduction.
Optimizing utility and privacy ratio is an NP-hard problem. Therefore, we are required to develop and design a model that can transform data such that no confidential
information gets revealed but must retain high utility. Domains like heuristics, border schemes, evolutionary, and linear/ integer programming have been explored to
achieve the goal. Heuristics are simple and fast however but have maximum side-effect and lowest data utility. Border-based schemes perform better than heuristics in
terms of utility but are computationally expensive and difficult to apply in real-time bound applications. Linear and integer programming schemes are computationally
exhaustive schemes but provide the most optimized
transformed data. Evolutionary schemes have been explored recently and are providing near-optimal solutions. Further, these schemes are computationally less
Privacy preserving sensitive exhaustive than border and exact schemes. In this project, we shall develop new models to transform the dataset. Further, these model will induce anonymity in the
information masking scheme
CPG 192 target datasets. This will help in transformation of data with better privacy, utility, and low side effects.
Lumina Link is a SaaS platform designed to empower B2B companies by personalizing website content for each visitor, thereby improving conversion rates and
reducing bounce rates. Lumina Link addresses these challenges by automating the personalization process, utilizing data integration, machine learning, and a visual
content editor to simplify user interaction. The platform integrates seamlessly with existing marketing tools and offers real-time personalization recommendations based
on visitor segmentation.
1. Maximize Conversions: Sharply increase the percentage of website visitors who convert into paying customers or take desired actions.
2. Data-Driven Optimization: Empower businesses to make informed website decisions based on deep customer insights, not guesswork.
Lumina Link
CPG 193 3. Effortless Personalization: Simplify the creation of personalized website experiences for different audience segments.
Algorithmic forex trading and back testing system, Problem statement: Making profits in forex markets is getting difficult due to its increased dependence on more and
Algorithmic forex trading and back more factors, Sol: To design and Implement a sophisticated and robust live 24/7 trading bot, the trading bot will utilise multithreading, work queues and advance
testing system
CPG 194 strategies backed by comprehensive back testing system.
Objective - Create a functional prototype of the smart helmet system integrating alcohol detection, helmet
Enhancing Motorcycle Safety with compliance verification, distance monitoring, accident detection, and emergency communication
Smart Helmet Technology
CPG 195 features.
Project Definition:
Developing a classroom attendance system using LiFi technology ensures that attendance is logged exclusively when students are physically present. The solution
involves deploying a mobile application for students to register attendance within the LiFi coverage area. Backend integration streamlines attendance management by
automatically recording attendance based on received roll numbers and class schedules. This approach enhances security in attendance marking, eliminates proxy
attendance, and simplifies administrative tasks for educational institutions.

Solution Proposed:
The proposed solution entails implementing LiFi technology to securely transmit data within classroom confines. Attendance marking will be restricted to the LiFi
coverage area through a user-friendly mobile application. Backend systems will store timetables and student information, enabling automated attendance tracking based
on roll numbers.

Outcome:
The implementation ensures precise attendance records, thereby enhancing accountability within educational settings. By automating attendance recording, the solution
LiFi-Based Optimized Attendance reduces administrative workload and minimizes errors. Additionally, it facilitates improved tracking of student participation, ultimately leading to enhanced educational
Tracking System
CPG 196 outcomes.
BAATCHEET is an AI-based emotional companion designed to combat social isolation and improve mental health. This innovative system processes audio input using
a local emotion recognition model to assess the user's emotional state. It then generates dynamic, empathetic responses in real-time, simulating a conversation with a
Baatcheet
CPG 197 supportive friend
Automated Automobile Surveillance System, an innovative campus surveillance system, brings about a paradigm shift in the realm of campus security and
administration. It intelligently identifies and monitors authorised vehicles—cabs, deliveries, faculty and student vehicles, as well as those transporting parents—via
sophisticated camera networks. In addition to overseeing the movements of vehicles, the system authenticates their intended destinations to ensure they arrive at their
designated areas. In addition, we are in the process of creating an online portal that is intuitive and enables effortless vehicle registration and monitoring. By
Automated Automobile Surveillance implementing this integrated approach, various benefits are realised, including improved campus safety, streamlined traffic flow, and valuable data insights. It
System
CPG 198 revolutionises the management of vehicular access on college campuses, guaranteeing a safe and effective setting for all participants.
Sleek Space : An AR Based Interior Experience immersive architectural visualization in our Unreal Engine VR building. Explore realistic environments, interact with interactive elements, and bring your
Decor Applications
CPG 199 designs to life in virtual reality."
1. Virtual reality tourism is like taking a computer-generated trip where you can see, hear, and feel like you're in a different place without actually traveling there.
2. Virtual reality existing nowadays are not interactive and customizable. So, these are the areas where we will be working on.
3. We have added the scope of customizing the environment for the user and making it socially interactive by integrating it with a social media platform.
V
CPG 200 irtual Reality Tourism Experience 4. Outcome of the project would be an environment which would incorporate VR.
Project Name: Automated Plant Growth System

Problem Statement: Busy individuals lack the time or expertise to maintain optimal growing conditions for their homegrown crops, leading to suboptimal yield and
plant health.
Target Audience: Home gardeners, Urban dwellers, individuals seeking fresh produce without a large time commitment.

Solution:
System Overview: An automated system using various sensors (soil texture, drainage, pH, air quality, weather, temperature) to gather data and AI-powered software
to analyze it. Based on the analysis, the system automatically:
1.Dispenses customized water and nutrient mixtures.
2.Applies targeted pesticides and insecticides (minimizing unnecessary usage).
3.Regulates artificial light exposure to simulate optimal sunlight conditions.

Outcome:
Impact on Plant Growth: Increased yield, healthier plants, improved resistance to disease and pests.
Automated Plant Growth System
CPG 201 User Benefits: Convenience, reduced need for constant monitoring and intervention, higher success rate in growing homegrown crops.
The primary problem addressed by CipherX is the lack of a comprehensive online proctoring solution that can ensure the integrity of online exams while also providing
a positive user experience. The scope of this project includes the development of a proctoring system that integrates advanced security features such as AI-driven
cheating prevention, facial recognition, and dynamic tab monitoring, with a user-friendly interface that is accessible to students worldwide.
The project aims to create a scalable, secure, and reliable platform that can handle large volumes of concurrent users, while also ensuring compliance with data privacy
regulations. The system will be designed to work across a wide range of devices and operating systems, making it accessible to all students, regardless of their technical
CipherX- The Era of Proctoring background
Excellence
CPG 202
 Capstone projeCts 2024

The project is an innovative online platform designed to provide career counselling based on the analysis of dropout trends in colleges and schools. The core of the
From Classrooms to Crossroads – An platform is a predictive model that uses responses from a detailed questionnaire to assess potential dropout risks and then provides tailored counselling through an AI-
initiative to Career Counselling along driven chatbot.
with Predictive Dropout Analysis
CPG 203
Randomization using Quantum.
CPG 204 We will develop a game which is based on unpredictable or random patterns. These random numbers will be generated with help of quantum.
Project Description:
We are developing a state-of-the-art safety helmet equipped with IoT technology tailored for workers in high-risk construction sites. The helmet will feature an array of
sensors aimed at enhancing communication and safety measures.

Project Problem:
In hazardous construction environments, traditional safety measures are often insufficient. Workers face risks that demand a more advanced and proactive safety
solution. Existing communication gaps and the lack of real-time data analysis pose significant challenges.

Project Solution:
Our solution involves integrating IoT devices into safety helmets to create a comprehensive safety system. These devices, including sensors, will monitor various
parameters in real-time, fostering improved communication and ensuring immediate responses to potential dangers. The collected data will be processed using AI
algorithms for insightful analysis.

Outcome:
S.H.I.E.L.D. - Safety Helmet The implementation of this IoT-enabled safety helmet will lead to a paradigm shift in construction site safety. Real-time monitoring and analysis of data will provide a
Integrated with Enhanced proactive approach to risk mitigation, significantly reducing accidents and improving overall worker well-being. The seamless integration of technology will
Learning and Detection
CPG 205 revolutionize safety standards in the construction industry.
The project aims to develop an AI chatbot capable of initiating and engaging in human-friendly conversations while also providing various home automation
functionalities. The chatbot will be trained to understand specific user intents related to tasks such as controlling home appliances (e.g., lights and fans), searching
Wikipedia for information, playing songs from YouTube, retrieving the current time, and setting reminders.
The proposed solution involves designing and implementing an AI chatbot system with components like Natural Language Understanding ,Dialog Management,
Reminder System, Integration with External APIs.
AI Home Assistant
CPG 206 Upon completion, the AI chatbot will offer users a convenient and efficient way to perform various tasks through natural language interactions.
Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition that
presents significant challenges in diagnosis due to its wide range of symptoms and varying
levels of severity. Traditional diagnostic methods, though effective, are often timeconsuming, subjective, and require extensive training. This project aims to develop a
mobile application designed to streamline the diagnostic process for ASD by integrating
standardized diagnostic tests, automating scoring, and providing real-time data processing.
The application leverages modern technologies, including Flutter for cross-platform
development, Firebase for secure data storage.. The primary objective is to create a userfriendly tool that allows psychologists to administer ASD diagnostic tests
efficiently, with
immediate feedback and reduced manual errors.
Through comprehensive investigative techniques—including descriptive, comparative,
and experimental methods—the project has validated the application’s ability to enhance
the accuracy and speed of ASD diagnosis. Initial testing with a group of psychologists has
shown promising results, indicating that the application can significantly reduce the time
required for diagnosis while maintaining or improving the accuracy of traditional methods.
The project concludes with recommendations for future enhancements, including the
integration of machine learning algorithms to further improve diagnostic accuracy and the
potential for global deployment to reach underserved communities. This application has
Enhanced Autism Detection: A the potential to become an invaluable tool in the early detection and diagnosis of ASD,
Mobile App for Psychologists
CPG 207 ultimately leading to better outcomes for individuals affected by the disorder.
This project introduces an effective method for forecasting Parkinson’s disease progression with limited data, utilizing the Movement Disorder Society Unified
Parkinson’s Disease Rating Scale (MDS-UPDRS) and the Accelerating Medicines Partnership-Parkinson’s Disease Kaggle dataset. Employing machine learning,
A Comprehensive Study on Deep including deep learning with the Multi-Layer Perceptron (MLP), the study demonstrates superior predictive performance across patient clusters. The transition to a
Learning Techniques for Parkinson's supervised format enhances performance, emphasizing the efficacy of deep learning in capturing intricate trends within limited data constraints and contributing to the
Disease Progression Forecasting
CPG 208 advancement of early detection and forecasting of Parkinson's disease progression.
AgroFinance is a project at the intersection of blockchain finance and agricultural supply chain management, tailored to empower Small and Medium-Sized Enterprises
(SMEs) in the farming sector. This platform seamlessly integrates decentralized finance (DeFi) protocols, providing farmers with secure and flexible financing options.
Through smart contract-powered lending processes, transparent and auditable transaction records, and a user-friendly interface, AgroFinance not only revolutionizes
financial inclusion but also incorporates an advanced supply chain management system. By enabling farmers to optimize their supply chains, AgroFinance aims to
AgroFinance
CPG 209 reduce costs, enhance traceability, and create a decentralized financial ecosystem that fosters sustainable growth within the agricultural community.
DEFINITION=Face detection is a computer vision task that involves identifying and locating human faces in digital images or video frames. This technology has
gained significant attention and practical applications in various fields, such as security, entertainment, healthcare, and more.
SOLUTION PRPOSED=Collect and annotate a diverse dataset.
Choose a suitable CNN architecture (e.g., SSD, YOLO).
Train the model on the dataset.
Evaluate performance metrics.
Deploy and monitor for real-world usage, considering ethical implications and continuous improvement
OUTCOME=1. Identity Verification: One of the primary outcomes is the ability to verify and authenticate the identity of an individual based on their facial features.
2. Access Control: Face recognition systems enable efficient access control by allowing or denying entry to secured areas based on recognized faces.
3. Law Enforcement and Criminal Identification: Face recognition assists law enforcement agencies in identifying individuals from surveillance footage or images.
FACE DETECTION PORTABLE DEVICE 4. Time and Attendance Tracking: In business and educational settings, face recognition systems facilitate accurate and efficient time and attendance tracking.
CPG 210
The Automated Targeting System for Human Vital Points Detection ,
cutting-edge solution designed to accurately locate and track crucial
anatomical landmarks on the human body. By harnessing the power of
laser guidance and dual-camera integration, this system offers precise
targeting and refined imaging capabilities.

The system identifies vital points on the human body with unparalleled
accuracy. This laser guidance serves as a reliable marker for
Lakshya - Automated Targeting pinpointing specific anatomical locations, facilitating precise
System
CPG 211 targeting in various applications.
 Capstone projeCts 2024

The goal of this project is to implement a system that manages seating within the library by providing real-time seat occupancy status at the entrance for students and
analytics for the management. Utilizing live CCTV footage, object detection techniques will be employed to ascertain seat occupancy status, providing a valuable tool
for students seeking available seating.

Objectives

Develop a system to extract real-time seat occupancy status from live library CCTV footage.
Show real time analytics (no. of people in the library, projected footfall at any give day/month)
Provide a seamless and efficient method for students to identify available seating without relying on RFID or pressure sensors.

Methodology

Data Collection and Preparation:

Video to Frames: The live library CCTV footage will be converted into individual frames for further analysis.

Data Preprocessing:
Techniques will be applied to address challenges posed by different lighting conditions, ensuring accuracy in seat occupancy detection.

Model training and prediction:

The system will refresh and update the screen displaying seat occupancy status every 60 seconds based on the processed data from the CCTV footage.

Data stored in database: The output data will be stored in a database which will be used for further analytics.

Advantages:

Time-Saving: Particularly beneficial during examination periods, where students can quickly locate available seats.
Improved Management: Offers a more efficient alternative to the existing RFID and pressure sensor-based systems.
Real-Time Seat Occupancy Status Ease of Implementation: No initial prerequisites or extensive setup required.
System for Library
CPG 212 Enhanced Library Management: Facilitates better utilization of library resources.
ATM plays an important role in modern economic activity. Unfortunately, it provides a convenient way for criminals to get illegal money. For safety reasons, each
ATM is equipped with a surveillance system to record customer's face information. When criminals use the ATM to withdraw illegal money, they usually hide their
faces with something to avoid the surveillance system recording their face information. Due to this the efficiency of the surveillance system is decreased. In this project
the process of ATM transaction will be finetuned so that a person can’t proceed in the transaction without removing the mask and helmet. With each transaction the
image of the person along with the transaction details will be stored.Primary goal is to secure ATM transaction processing. If we achieve this, no customer will be
Mask and Helmet Recognition for taken advantage of, and there won't be any misunderstandings between the bank and its clients. And if we have an accurate count of the number of customers and the
ATM Security
CPG 213 precise person doing the transaction, then if a problem arises in the future, we can use these as a reference to find the issue.
security and streamline parking management. This system utilizes advanced technology to scan and record the number plates of vehicles entering and exiting the
campus, while also tracking the duration of their stay.

**Key Features:**

1. **License Plate Recognition (LPR):** The system employs state-of-the-art LPR technology to automatically capture and recognize the license plates of vehicles
entering the campus. This enables seamless identification and tracking of vehicles without manual intervention.

2. **Real-time Monitoring:** A centralized desktop software provides real-time monitoring of vehicle movements within the campus. Authorized personnel, including
students and faculty members, can access this software to register their vehicles and view their parking duration.

3. **Vehicle Registration:** Students and faculty members can register their vehicles through the desktop software by providing relevant details such as license plate
number, vehicle model, and owner information. This registration process ensures accurate tracking and management of authorized vehicles.

4. **Duration Recording:** The system records the duration for which each registered vehicle remains inside the campus premises. This information helps in analyzing
parking patterns, optimizing space utilization, and ensuring campus security.

5. **Alerts and Notifications:** The system can generate alerts and notifications for various events, such as unauthorized vehicle entry, parking violations, or vehicles
exceeding the allowed duration. This proactive approach enables prompt action and ensures compliance with parking regulations.

6. **Integration with Access Control:** For enhanced security, the system can be integrated with existing access control mechanisms to restrict entry and exit based on
vehicle registration status. Only authorized vehicles are granted access to the campus premises.

7. **Data Analytics and Reporting:** The system generates comprehensive reports and analytics based on vehicle movement data, parking duration, and other relevant
metrics. These insights aid in decision-making, capacity planning, and optimizing parking facilities.

**Benefits:**

- **Enhanced Security:** By accurately tracking vehicle movements and enforcing parking regulations, the system enhances overall campus security.
Campus Parking System
CPG 214 - **Efficient Parking Management:** Real-time monitoring and data-driven insights enable efficient utilization of parking spaces, reducing congestion and improving
In today's urban environments, traffic congestion and safety concerns pose significant challenges to efficient transportation systems. To address these issues, our project
introduces a comprehensive Smart Traffic Management System that leverages cutting-edge technology to optimize signal control and enforce traffic rules effectively.

The core features of our system include dynamic traffic signal optimization based on real-time traffic density analysis and intelligent detection and tracking of traffic
rule violations. By integrating artificial intelligence, deep learning, machine learning algorithms, and computer vision techniques, our system continuously monitors
traffic conditions and adjusts signal timings accordingly to ensure smooth traffic flow and minimize congestion.

Additionally, our system incorporates a robust rule enforcement mechanism that detects and traces traffic rule breakers. Utilizing video surveillance and AI-powered
Next-Gen Traffic Control
CPG 215 analytics, the system identifies violations such as red light running, illegal turns, and speeding, enabling authorities to take appropriate actions swiftly.
This project involves creating a Handwritten Text Recognition and Analysis system using OCR, Deep Learning, and CNN. It aims to accurately transcribe and
Handwritten OCR
CPG 216 understand handwritten text from images, supporting multiple languages and enhancing accuracy through advanced deep learning techniques.
 Capstone projeCts 2024

This proposal outlines the development of a multilingual chatbot aimed at

assisting farmers in India, addressing their needs in both Punjabi and Hindi
languages. This innovative tool will leverage natural language processing
(NLP), artificial intelligence (AI), and object recognition to bridge the
information gap and empower farmers, particularly those facing literacy
challenges.
2. Problem Statement:
• Limited access to agricultural information in Punjabi and Hindi,
creating a barrier for many farmers, especially illiterate and semi
literate ones.
• Difficulty for farmers to access expensive research, expert advice, and
government schemes, impacting decision-making and farm
productivity.
• Lack of timely and readily available information on weather, diseases,
and market trends.
3. Proposed Solution:
We propose developing a Punjabi-Hindi Language Learning Model (LLM)
chatbot with the following features:
• Multilingual Text-to-Speech and Speech-to-Text: Enables seamless
interaction in both Punjabi and Hindi, overcoming literacy barriers.
• AI-powered Information Retrieval: Provides quick and easy access to
agricultural knowledge, research findings, and expert advice in
preferred languages.
• Disease Detection using Object Recognition: Offers early disease
identification through mobile phone camera images.
• Weather Forecast Predictions and Alerts: Delivers localized weather
data and timely alerts in chosen languages.
KhetMitra: Your AI-powered • Government Scheme Updates: Keeps farmers informed about
Agricultural Assistant in Punjabi and relevant government initiatives and programs.
Hindi
CPG 217 • Market Trend Analysis: Provides insights into commodity prices and
NA
CPG 218 NA
Introduction: RetinaVision is an innovative project aimed at AI and ML to develop a sophisticated model for the automated diagnosis of diabetic retinopathy. By
analysing retina images, our system aims to detect signs of diabetic retinopathy, potentially enabling early detection and treatment.
Key Objectives:
Diagnosis Accuracy: Develop a highly accurate AI/ML model capable of detecting diabetic retinopathy in retina images. Early Detection: Focus on early detection to
facilitate timely medical intervention and prevent complications. Accessibility: Create a user-friendly website interface where individuals can easily upload their retina
RetinaVision: Automated Diabetic images and receive a prompt assessment of diabetic retinopathy probability. Scalability: Ensure the system can handle a large volume of image submissions
Retinopathy Diagnosis
CPG 219 concurrently without compromising performance. Privacy: Implement robust data privacy measures to safeguard users' sensitive medical information.
Project Description:

Our project, "EchoSign," will address communication gaps for individuals using American Sign Language (ASL). The sensory glove, equipped with flex, contact
sensors, and an accelerometer, will capture precise hand movements. Leveraging Principal Component Analysis, the glove will be going to learn to recognize ASL
gestures, translating them into alphabets/phrases in real-time. The device will establish a Bluetooth link with an Android phone, displaying received letters and words
while converting them into audible speech.

Solution Proposed:

EchoSign's innovative solution will seamlessly integrate cutting-edge sensor technology with machine learning. The glove's sensors will efficiently capture nuanced
hand expressions, and PCA will ensure accurate gesture classification. The seamless Bluetooth connection to an Android phone will facilitate real-time translation,
displaying interpreted text and converting it into spoken words. This comprehensive approach will establish a natural and immediate communication bridge between
ASL users and those relying on spoken language.

Outcome:

EchoSign Translating Glove will empower the deaf and hard of hearing by seamlessly translating ASL gestures into spoken language. The glove's real-time recognition
and translation capabilities will ensure prompt and accurate communication. The integration with an Android phone will enhance the user experience, displaying
EchoSign : Sign Language Translation translated text and converting it into clear, audible speech. Our overarching goal is to contribute to a more inclusive society, dismantling communication barriers and
through Sensory Glove
CPG 220 fostering effortless interaction between ASL users and those unfamiliar with sign language.
E-rickshaw and smart bike app for Revolutionize campus mobility with our innovative project integrating traditional rickshaws and cutting-edge smart bike technology. Our app seamlessly connects users
campus
CPG 221 with nearby rickshaws or bikes, providing efficient and eco-friendly transportation solutions. Navigate campus effortlessly
This project presents an advanced fire detection system utilizing the YOLOv8 object detection model for real-time identification of fire and smoke from various input
sources, such as live camera feeds, pre-recorded videos, and static images. The system processes each input to detect fire hazards and triggers automated email alerts
with timestamped snapshots to notify users. The model is pre-trained with a dataset optimized for fire and smoke detection, ensuring high accuracy in real-time
applications. Additionally, the system includes a user-friendly graphical interface, offering options for video analysis, image inspection, and live camera feed
monitoring. Optimizations such as frame-skipping for video and camera feeds ensure efficient processing, while maintaining the system's performance.
This system is designed to support fire safety monitoring and emergency response applications, where rapid detection and alerting are critical. Its modular design, real-
Fire Detection Using YOLO
CPG 222 time processing capabilities, and automated alert mechanism make it a reliable solution for safeguarding environments prone to fire hazards.
This project includes a physical chess board with a opponent(Robot). It may include a screen which will show the possible options . Opponent will have two options
one will include fully automated and other will include assisted help in which it help beginner who might not understand chess much. Robotics arm and motion detector
Automated Robotic Chess Player
CPG 223 and other
Project sensors like scanner might be used.
Definition:
CipherChat is a secure messaging application that leverages a combination of blockchain technology and cryptography to ensure the privacy, integrity, and authenticity
of user communications. Traditional messaging applications rely on centralized servers, posing risks such as single points of failure and vulnerabilities to unauthorized
access. In contrast, CipherChat aims to address these concerns by decentralizing message delivery and implementing end-to-end encryption on a blockchain network.

Solution Proposed:
CipherChat utilizes blockchain technology to deliver encrypted messages between users, ensuring that only the sender and receiver can decrypt and read the messages.
Smart contracts are employed to verify user identities, preventing impersonation and unauthorized access. Messages are stored on the blockchain, making them
immutable and tamper-proof, thereby ensuring message integrity. Additionally, the transparent and timestamped nature of the blockchain provides an audit trail for
message verification, enhancing accountability and trust.

Outcome:
By implementing CipherChat, users will benefit from enhanced privacy, security, and control over their communications. The decentralized nature of the application
eliminates single points of failure, ensuring high reliability and availability. End-to-end encryption guarantees that messages remain confidential and can only be
accessed by intended recipients. Furthermore, the immutability of messages on the blockchain and the use of smart contracts for user verification provide additional
CipherChat:-Where Blockchain layers of security, trust, and accountability. Overall, CipherChat offers a robust and secure messaging solution that addresses the shortcomings of existing
meets Cryptography
CPG 224 communication models.
MyCare: Automatic Pill Dispenser introduces a cutting-edge solution for efficient medication management. By seamlessly integrating hardware components like
microcontrollers, motors, and sensors with sophisticated software functionalities, the system ensures precise and timely dispensing of medication doses. Key features
such as timed dispensing, interactive reminders, refill alerts, and dose quantity checks enhance user control and monitoring, promoting adherence to medication
schedules. Additionally, innovative additions such as voice control, biometric authentication, and integrated display/speaker with loved one's image/voice elevate the
user experience, fostering improved communication and ease of use. Ultimately, this project aims to revolutionize medication adherence and healthcare outcomes
MyCare: Automatic Pill Dispenser
CPG 225 through advanced technology and user-centric design.
 Capstone projeCts 2024

In the realm of university campuses, the demand for a seamless and secure alternating commuting experience is paramount in fostering sustainable campus mobility.
Traditional rental methods often grapple with issues of accessibility and real-time tracking, subsequently giving rise to security concerns and are also time wasting.
This project embarks on a mission to redefine campus transportation by introducing an all-encompassing solution. This includes a centralized mobile application, smart
bicycle locks driven by IoT technology, and an efficient e-rickshaw booking system , all geared towards creating a more sustainable and eco-friendly campus
Sustainable Optimized Links for environment.An integrated and user-centric solution that not only simplifies bicycle rentals and e-rickshaw bookings on campus but also contributes to a sustainable
Alternative Commuting Experience
CPG 226 and eco-friendly campus ecosystem.
To provide advance warnings to nearby vehicles about adverse weather conditions and to promptly share information about road accidents,
EyeOnRoad: Vehicle-to-Vehicle Dedicated to transforming road safety through cutting-edge V2V communication technology.
Communication for Road Safety
CPG 227 It's revolutionizing driving experiences and raising the bar for traffic efficiency and safety standards.
This project aims to create prediction tools incorporating computational models and algorithms capable of accurately predicting Class I MHC restricted T cell epitopes.
By leveraging machine learning techniques, bioinformatics methodologies, and vast datasets of experimentally validated epitopes, the project seeks to enhance our
ability to identify antigenic peptides that can bind to Class I MHC molecules and elicit robust T cell responses.

Key components of the project include:

1. Data Collection and Curation: Gathering datasets of experimentally verified Class I MHC restricted T cell epitopes.
2. Feature Extraction and Selection: Identifying relevant features and characteristics of epitopes and MHC molecules.
3. Machine Learning Model Development: build models for classifying epitopes based on their binding affinity to Class I MHC molecules.
Development of prediction tools for 4. Model Evaluation and Validation
Class I MHC restricted T cell epitope 5. Tool Development and Deployment: Implementing user-friendly web-based interfaces to make the prediction tools accessible
CPG 228
Smart Farming combines advanced AI techniques to revolutionize fruit cultivation. By leveraging Convolutional Neural Networks (CNNs) for fruit identification and
Smart Farming
CPG 229 Extreme Learning Models for weather, soil, and water predictions, this innovative approach enhances resource efficiency and yields.
and surface imperfections. This project seeks to enhance efficiency and accuracy in pomegranate sorting, reducing labour costs and minimizing waste in the agricultural
Crop Quality Management
CPG 230 industry.
Description:
Deepfake detection identifies manipulated or synthetic media content using machine learning algorithms and computer vision techniques to detect anomalies in facial
and body movements and other visual.
It includes pre-trained models such as

1. MTCNN (Multi-Task Cascaded Convolutional Networks) for face detection and alignment, and
2. InceptionResnetV1 for detecting whether an image is fake or real.

What we are going to deliver:

* Trained Model: A machine learning model trained on a dataset of both real and fake images to accurately detect and classify deepfake images.

* Detection Software/Application: A software or application that utilizes the trained model to detect deepfake images. This could be in the form of a standalone
desktop application, a web-based tool, or even integration into existing platforms

* Documentation: Detailed documentation explaining the methodology used, the architecture of the model, instructions for deploying and using the detection software,
and guidelines for interpreting the results.

* Evaluation Metrics: Analysis and reporting of the performance metrics of the model, such as accuracy, precision,etc., to assess its effectiveness in detecting deepfakes.

Verity : A DeepFake Detection Engine * Training Data: providing the dataset used for training the model, along with any necessary preprocessing steps and annotations.
CPG 231
MediLocker revolutionizes medical record management by offering a centralized database for users to securely store and maintain their records. With features like
direct prescription uploads from healthcare professionals, consent-based sharing, QR code-enabled information exchange at medical facilities, and seamless sharing of
reports with insurance companies, it ensures efficiency and credibility in healthcare interactions. Additionally, a Chatbot and community section provide support for
MEDILOCKER
CPG 232 minor health queries, making it a comprehensive solution for individuals seeking accessible and reliable healthcare management.
In addressing the declining health of elderly patients, an imperative focus lies on continuous care, vigilant monitoring, and timely disease detection to maximize
treatment efficacy and bolster chances of recovery.
We propose a Social Assistant System that can detect diseases in the elderly by using data from multiple sensors(ECG, EEG, etc) to facilitate early intervention and
treatment by healthcare professionals. It also monitors vital signs of the patient like body temperature, blood oxygen level, heart rate, etc. which can be accessed by the
patient and the medical caregivers through an app.
The app alerts medical caregivers in case any parameters deviate from normal ranges. It also sends timely reminders to the patient to take their medication and suggest
remedies based on the disease detected.
Future prospects for this project include an enhanced user experience through the integration of natural language processing (NLP) and voice chat functionalities to
augment the capabilities of the app.
Social Assistant for the Elderly
CPG 233 We aim to build a software that in the future can be deployed on a social robot like NAO by the concerned authorities.

The aim of this project is to develop a smart mental health monitoring system that utilizes embedded sensors to collect relevant physiological and behavioral data from
individuals. This data will then be processed using machine learning algorithms to predict and analyze patterns associated with mental health conditions, enabling early
detection, intervention, and personalized support.

Proposed Solution:

The proposed solution involves the following components:

Embedded Sensor Network: Deploy sensors such as accelerometers, heart rate monitors, skin conductance sensors, and sleep monitors in wearable devices or smart
home environments to continuously collect data related to physical activity, heart rate variability, sleep patterns, and stress levels.

Data Collection and Integration: Establish a data collection pipeline to aggregate sensor data in real-time. This may involve designing firmware for sensor devices to
collect and transmit data securely to a centralized server or cloud platform for further processing.

Machine Learning Models: Develop machine learning algorithms capable of analyzing the collected data to detect patterns indicative of various mental health
conditions such as anxiety, depression, stress, or bipolar disorder. These models may utilize techniques such as supervised learning for classification tasks or
unsupervised learning for anomaly detection.

Predictive Analytics: Implement predictive analytics algorithms to forecast potential mental health crises or episodes based on historical data patterns. This could
involve time-series analysis, anomaly detection, or deep learning approaches to identify early warning signs and trigger preemptive interventions.

User Interface and Feedback: Design a user-friendly interface for individuals to access their mental health data, receive personalized insights, and access resources or
support services as needed. This interface may be web-based, mobile applications, or integrated with existing health monitoring platforms.

Privacy and Security: Ensure robust privacy and security measures are in place to protect sensitive health data collected from individuals. This includes data encryption,
user consent mechanisms, and compliance with relevant regulations such as HIPAA or GDPR.

Smart Mental Health Monitoring Outcome:

System
CPG 234
 Capstone projeCts 2024

Definition: Exercise Accuracy Detection Using Smart Wearables

Solution Proposed: We aim to utilize data from two smart wearables to assess the accuracy of exercise positions. Traditional methods of assessing exercise form often
rely on subjective observations or costly motion capture systems. Challenges include synchronizing wearable data, developing analysis algorithms, and validating
accuracy. This exercise uses smart wearable data to accurately track and evaluate a user's performance against fitness goals. The project aims to develop a robust and
reliable solution for assessing exercise positions using smart wearables, ultimately improving the effectiveness and safety of fitness training programs.

Project outcomes: The project will deliver a comprehensive solution to assess exercise positions using smart wearables. This means users can rely on these devices to
track and evaluate their performance against fitness goals accurately. With smart wearables, users can effortlessly monitor and refine their exercise form without
constant human guidance, making fitness training more accessible. Plus, they'll receive instant feedback on their exercise positions, allowing them to adjust in real-time
Exercise Accuracy Detection Using and work towards their fitness objectives more effectively. Ultimately, this project aims to enhance training programs, making them safer and more efficient and
Smart Wearables.
CPG 235 empowering users to achieve their fitness targets while minimizing the risk of injury.
This capstone project explores the design and development of a solar-powered, mobile app-controlled oil skimmer boat, created to tackle the pressing environmental
issue of oil spills in aquatic environments. The project integrates renewable energy, robotics, and wireless communication technologies to offer a sustainable and
Solar-Powered Mobile App efficient solution for oil
Controlled Oil Skimmer Boat
CPG 236 spill remediation.
This project involves the development of an autonomous buggy system designed to identify and differentiate between leaves and plastic items amidst various obstacles.
The system leverages sophisticated machine learning models and image processing technologies to achieve accurate classification. By processing visual data, the
system effectively segregates leaves from plastic items, thereby enhancing the efficiency of environmental monitoring and waste management applications. The
Grass-O-clean
CPG 237 integration of these technologies ensures precise detection and sorting, demonstrating advanced capabilities in autonomous navigation.
3D reconstruction of 2D dental In this project, our objective is the advancement of a cost-effective and highly accurate methodology for developing three-dimensional models of dental structures.
images
CPG 238 Leveraging deep learning principles, we intend to achieve this by utilizing 2D cone-beam computed tomography (CBCT) images as input data.
The imminent introduction of the Automated Grading and Secure Transcript Management System, "ScanScore," represents a pivotal moment in the evolution of
academic assessment. This groundbreaking system is poised to revolutionize the conventional grading process by harnessing the power of automated scanning
technology, facilitating the seamless transfer of individual question marks to a centralized database. The inherent elimination of manual data entry not only anticipates a
substantial reduction in errors but also foresees the expeditious progression of the grading process, marking a new epoch of efficiency in academic evaluation. In Indian
education, particularly within centralized examination bodies like the Central Board of Secondary Education (CBSE), where the sheer volume of answer sheet
necessitates a more streamlined approach, ScanScore emerges as a timely solution to meet the pressing needs of efficiency and accuracy.
ScanScore distinguishes itself with a robust security infrastructure that prioritizes the confidentiality and integrity of academic records. This commitment to security
becomes increasingly crucial within the context of major examination bodies like CBSE, where the protection of student information is paramount. Noteworthy is the
fact that ScanScore diverges from traditional approaches, eschewing the development of a graphical user interface. Instead, educators will be presented with a user-
friendly Excel sheet, ensuring facile access to review and validate recorded grades. It provides a simple and accessible method for educators to navigate and manage
academic assessments.
Moreover, ScanScore's automatic sheet unwinding mechanism minimizes manual effort associated with handling answer sheets, saving valuable time for educators and
enhancing grading efficiency. This automation, particularly impactful for large entities, is coupled with a simplified Excel output, making ScanScore a forward-thinking
solution for evolving assessment needs in the Indian educational landscape. Its user-focused approach positions it as an instrumental tool for streamlining processes,
Smart Scan
CPG 239 enhancing accuracy, and embracing the challenges of scale within India's diverse educational ecosystem.
The proposed work integrates illuminance enhancement with depth-guided refinement and employs Generative Adversarial Networks (GANs) for effective denoising.
DED-GANet: Decomposed, Enhanced The culmination of these techniques shows promising advancements in low-light image enhancement. Through extensive experimentation and thorough evaluation, the
with Depth, and GAN-Infused Image project aims to quantify and qualitatively assess the efficacy of DED-GANet, hoping to position it as a game-changer in the realm of low-light image processing. We
Enhancement
CPG 240 aim to deliver a webapp capable of taking low light images as input from the users and output enhanced version of the same image.
Learning assistant for blind
CPG 241 A system designed to help blind children to learn braille
Multimodel on prediction Multimodel on early prediction of dementia using voice and retina oct scans(Parameters can be altered on the basis of further study of the group as well as consultation
of dementia.
CPG 242 from the mentors)
Person drowning detection using deep learning techniques.
1. Create own dataset.
2. Study various techniques for activity recognition.
Person Drowning Detection system 3. Apply these techniques to detect the person who is drowning and create a early alarming point.
CPG 243
The basic idea is to implement a Legal AI solution which will be used to cater around in providing judgement for the cases that go around in courts. Also, solution will
target law students to prepare law by getting useful simplification and extracting on point laws regarding different situations.
Legal AI Solution: Enhancing Legal Implementation will be based on Natural Language Processing, initially starting with collecting supervised and unsupervised data around all laws and then training
Processes through NLP
CPG 244 models based on RNNs, LSTM, transformers and transfer learning. Lastly create a LLM chat model and use reinforcement learning to get better user experience.
Project Definition: these days there is a lot of increase in roadside garbage. So, we are making smart cameras using machine learning to detect such garbage wastes.

Solution Proposed: detecting the garbage thrown at which place and by which vehicle.

Smart Garbage Detector

CPG 245 Outcome: reduction in roadside garbage throwing activities.
The objective of this task is to develop a reliable algorithm or model that can accurately identify
and differentiate between individuals based on their gait patterns. The algorithm should be able
to handle variations in clothing, footwear, walking surface, lighting conditions, and other
factors that may affect the gait pattern. It should be designed to work with one or multiple
videos of individuals walking and produce accurate results consistently.
Solution : The ambitious endeavor to advance human identification through gait analysis underscores the
inexhaustible potential of biometric innovation. In a world characterized by dynamic
technological progress, gait recognition emerges as a hallmark of cutting-edge identification
methodologies. This project, with its multidimensional exploration of gait analysis, stands as
a testament to the fusion of pioneering research and practical application. From algorithmic
intricacies to ethical considerations, the journey encapsulates the essence of modern biometric
advancements, poised to redefine the future landscape of human identification.

Human identification using gait Goal : The central objective of this ambitious endeavor is to develop an innovative algorithm capable
analysis
CPG 246 of accurately identifying individuals based on their intricate and individualistic gait patterns
Objective: Develop an advanced hospital appointment system
Key Features:
- RFID and AI integration for real-time doctor tracking
- Online appointment booking for patients
- Elimination of receptionists and paper-based processes
- AI-driven slot allocation for optimized scheduling
Hospital Scheduling System
CPG 247 - Enhanced efficiency and reduced wait times
Increasing the performance of Variational AutoEncoders by introducing Reversed AE (RA) using additional loss term, to generate pseudo- healthy reconstructions and
localize various brain pathologies.
Domain Independent AutoEncoders - 2. Developing a domain-independent deep learning model that is compatible with different Brain Tumor datasets across different medical imaging modalities (eg. X-
RCPG 1 for Anomaly Detection ray, MRI)
Arteriovenous fistula (AVF) formation is critical for establishing vascular access in hemodialysis patients. The success of AVF dramatically depends on selecting an
Intelligent Recommendation System optimal site, considering factors such as vessel diameter, blood flow, and long-term patency. The challenge lies in identifying the most suitable veins and arteries to
for Arteriovenous fistula Site enhance the overall efficacy and longevity of vascular access for hemodialysis patients. Making the recommendation system to find the best site for vascular access by
RCPG 3Selection testing different possible veins and arteries.
Exploring ideal watermarking techniques using Deep Learning for Medical Images.
Deep learning based watermarking Main criteria's will be to maintain a balance among the prime watermarking parameters including Robustness, Imperceptibility, withstanding attacks, and minimization
RCPG 4for e-Healthcare applications of Noise and Errors.
Quantum computing offers immense potential but poses significant algorithm design and optimization challenges. There is a need for novel algorithms that exploit the
Algorithm Design and Optimization unique properties of quantum systems to solve computational problems efficiently. This project aims to develop optimization techniques specific to quantum
RCPG 5Techniques for Quantum Computing algorithms, which is crucial to improve performance, reduce error rates, and enhance scalability.
Quantum computing offers immense potential but poses significant algorithm design and optimization challenges. There is a need for novel algorithms that exploit the
Algorithm Design and Optimization unique properties of quantum systems to solve computational problems efficiently. This project aims to develop optimization techniques specific to quantum
RCPG 8Techniques for Quantum Computing algorithms, which is crucial to improve performance, reduce error rates, and enhance scalability.
 Capstone projeCts 2024

Smart Breath Analyzer for TB

RCPG 9detection Research Project for detection of mycobacterium tuberclosis using qcm sensors by analysing VOC content in breath
Systematic Approach to Enhance Signals from the receiver of the antennas for remote sensing and astrophysical application often suffer from noise introduced by system devices, particularly in
Signal Quality and Noise Reduction transmission wires, compromising signal quality and reliability. The challenge lies in minimizing this unwanted noise to ensure optimal signal integrity and enhance the
in Remote Sensing and Astrophysical overall signal quality. This project aims to develop an efficient demodulation algorithm to mitigate noise, focusing on precise noise quantification and suppression. The
Applications
RCPG 10 expected outcome includes a notable improvement in signal quality and enhanced communication reliability.