OBJECT DETECTION

A PROJECT REPORT

Submitted by

HARIKRISHNA PANCHAL

210340131006

In partial fulfillment for the award of the degree of

BACHELOR OF ENGINEERING

in

Computer Science and Engineering

Narnarayan Shastri Institute of Technology – Jetalpur

Gujarat Technological University, Ahmedabad

April, 2025
 Narnarayan Shastri Institute of Technology
Jetalpur, Gujarat

CERTIFICATE
This is to certify that the project report submitted along with the project entitled Object
Detection has been carried out by Harikrishna Panchal under my guidance in partial
fulfillment for the degree of Bachelor of Engineering in Computer Science and Engineering, 8 th
Semester of Gujarat Technological University, Ahmedabad during the academic year 2025.

Prapti Kachhia Nimisha Modi

Internal Guide Head of the Department

GTU COMPLETION
 Narnarayan Shastri Institute of Technology

Jetalpur, Gujarat

DECLARATION

I hereby declare that the Internship / Project report submitted along with the Internship / Project

entitled Object Detection submitted in partial fulfillment for the degree of Bachelor of

Engineering in Computer Science and Engineering to Gujarat Technological University,

Ahmedabad, is a bonafide record of original project work carried out by me at CODING

CLOUD PVT. LTD. under the supervision of Harsh Patel and that no part of this report has been

directly copied from any students reports or taken from any other source, without providing due

reference.

Name of the Student Sign of Student

HARIKRISHNA PANCHAL ____________
TeamID:736478 Object Detection

ACKNOWLEDGEMENT
I have taken efforts in this project. However, it would not have been possible without the kind

support and help of many individuals and organizations. I would like to extend my sincere

thanks to all of them. I am highly indebted to Narnarayan Shastri Institute of Technology for

their guidance and constant supervision as well as for providing necessary information regarding

the project also for their support in completing the project. I would like to express my gratitude

towards my parents & members who offer their kind cooperation and encouragement which help

me in completion of this project. I would like to sincerely acknowledge and express my gratitude

to all those who have directly or indirectly been instrumental in preparing this report.

GTU i NSIT
TeamID:736478 Object Detection

ABSTRACT
This project focuses on implementing a real-time object detection system using

OpenCV and MediaPipe. The system is designed to detect and track objects,

including hands, faces, and other predefined objects, in live video input. By

leveraging computer vision techniques, it provides an interactive experience wher

users can interact with a visual marker (dot) on the screen. The system tracks user

movements, allowing interaction through gestures and time-based scoring. This

prototype serves as a foundation for future advancements, such as attendance

systems, automated surveillance, and traffic violation detection. The ultimate goal

is to enhance real-world applications using AI-driven object tracking, improving

automation, security, and user-interaction.

List of Figures

Object Detection lifecycle x

Flow-Chart x

Use case Diagram x

Developer module x

Data-science engineer module x

GTU ii NSIT
TeamID:736478 Object Detection

List of Abbreviations

AI Artificial Intelligence

ML Machine Learning

DL Deep Learning

CV Computer Vision

NN Neural Network

NLP Natural Language Processing

GTU iii NSIT

TeamID:736478 Object Detection

Table of Contents

Acknowledgement i
Abstract ii
List of Figures iii
List of Abbreviation iv
Table of Contents v
Chapter-1 Overview of the company 1
1.1 History 1
1.2 Different product/scope of work 1
1.3 Capacity of Department 2
Chapter-2 Production Management 3
2.1 Work being carried out in each department 3
2.1.1 Web solutions 3
2.1.2 App Development 3
2.1.3 Digital Marketing 3
2.1.4 Software Development 4
2.2 List the technical specification of major equipment used in each department 4
2.2.1 App Development 4
2.2.2 Website Development 4
Chapter-3 Internship Management 5
3.1 Internship summary 5
3.2 Purpose 5
3.3 Objective 5
3.4 Scope 5
3.5 Technology and literature review 6

GTU iv NSIT
TeamID:736478 Object Detection

3.5.1 Yaml 6
3.5.2 JavaScript 6
3.5.3 Jenkins 6
3.5.4 Docker 7
3.5.5 Docker compose 7
3.5.6 Github 7
3.5.7.AWS 8
3.6 Internship planning 8
3.6.1 Internship development approach and justification 8
3.6.2 Internship Effort and Time, Cost Estimation 10
3.7 Internship scheduling 11
Chapter-4 System Analysis 12
4.1 Study of current system 12
4.2 Problems and weakness of current system 12
4.3 System feasibility 12
Chapter-5 System Design 15
5.1 Flowchart 15
5.2 Use case Diagram 16
Chapter-6 Implementation 17
6.1 Implementation of Development Automation 17
6.2 Modules Specification(s) 18
6.3 System Interface 19
6.3.1 Github page 19
6.3.2 Jenkins page 20
6.3.3 Gmail page 22
6.3.4 Home page 23
6.3.5 Edit page 23

GTU v NSIT
TeamID:736478 Object Detection

Chapter 7 Conclusion and discussion 25

7.1 Summary of internship 25
7.2 Dates of continuous evaluation 25
7.3 Future enhancement 26
7.4 References 26

GTU vi NSIT
TeamID:736478 Overview of company

CHAPTER-1 OVERVIEW OF THE COMPANY

1.1 HISTORY

Coding Cloud Pvt. Ltd India based company. With more than 10 years of practical experience we
bring the experts in following Technologies:

● Python
● Data Science
● Data Analytics
● Machine Learning
● Business Analytics
● Software Testing
● Java Full Stack
● PHP Web Development
● Android

We have envisaged providing solutions to every IT-related problem in the most cost-friendly
way. With this noble vision, we have expanded globally giving innumerable IT solutions to our
widely existing and growing client and customer. Coding Cloud Pvt. Ltd. is receptive to new
ideas and promises help and support anytime and anywhere.

1.2 DIFFERENT PRODUCT/ SCOPE OF WORK

We are glad to realize complex web applications based on your design and specification. We are
experienced long-distance runners as most of the projects we have worked on are long-term and
usually lasts from 3 to 12 months. Our developers are constantly improving their qualification
levels. We’ve managed to organize our own learning system within our team. It helps our
developers increase their knowledge and always be up to date with innovative technologies and
be ready to apply them.

GTU 1 NSIT
TeamID:736478 Management
Overview ofthe company

1.3 CAPACITY OF DEPARTMENT

Vision: To become The Best IT Solutions and Services platform by giving optimal solutions to
Businesses and Educational Institutions for their consistent growth.

Mission: To provide such software solutions that are functional, reliable, maintainable and Cost-
friendly to our existing and growing client and customer base. To consistently cater to their
growing needs for an optimal solution, ensuring excellent support and service platform to give a
Hassle-free experience in achieving their dreams.

GTU 2 NSIT
TeamID:736478 Management
CHAPTER-2

PRODUCTION MANAGEMENT

2.1 WORK BEING CARRIED OUT IN EACH DEPARTMENT

In our Company, there are mainly four departments working like , App Development, Digital
Marketing and Software Development.

2.1.1 Web solution

The Web Solutions department at Coding Cloud is dedicated to helping businesses create a
strong online footprint. Our services include:Website Design and Development

● Custom Website Development

● Ecommerce Development
● Creative Graphics Design

2.1.2 Mobile Application Development

● Multi-Platform Apps: Developing responsive applications that provide a unified

experience across different devices.
● Windows-Based Solutions: Crafting optimized windows applications that cater to
business and user needs.
● Hybrid Mobile Apps: Blending web and native app technologies to build interactive and
feature-rich applications.

2.1.3 Digital Marketing

The Digital Marketing division focuses on strategies to maximize online visibility and audience
engagement. Our expertise includes: SEO Optimization, Social Media Promotion, Reputation
Management, Content Writing Service, etc.

GTU 3 NSIT
TeamID:736478 Management
2.1.4 Software Development

Software development involves various steps. Our team of experts takes care of all processes to
deliver functional and robust software. Firstly, we understand your requirements and then start
development. Our team also explores the scope for suggestions and improvements. After your
approval, we implement the suggestions for better performance of your software.

2.2 LIST THE TECHNICAL SPECIFICATIONS OF MAJOR EQUIPMENT

USED IN EACH DEPARTMENT

2.2.1 App Development

Our technical competency for undertaking application development contracts are listed below:

● Java / J2EE / J2ME Programming.

● Database (MongoDB, SQL Server, MySQL 5.0 and Oracle).
● ASP/PHP Programming.
● Visual C++ / C++ Programming.
● Asp.net / VB.net / C# Programming.

2.2.2 Website Development

● Javascript
● React JS
● Node JS
● HTML/CSS
● ANGULAR JS

GTU 4 NSIT
TeamID:736478 Internship Management
CHAPTER-3

INTERNSHIP MANAGEMENT

3.1 PROJECT / INTERNSHIP SUMMARY

I have done my internship at Coding Cloud Pvt. Ltd., Ahmedabad. The internship was for
duration of about 12 weeks and it involved a lot of learning activities all around. My internship
was based on Data Science Technology.

3.2 PURPOSE

The purpose of this project is to develop a real-time object detection system using OpenCV,
enabling machines to identify and classify objects instantly from live video feeds or images. This
system aims to enhance automation and efficiency in various industries by providing accurate
and fast object recognition. By leveraging OpenCV’s powerful computer vision techniques and
deep learning models, the project seeks to achieve low-latency detection while maintaining high
accuracy. The integration of real-time processing allows for applications in security surveillance,
autonomous navigation, traffic monitoring, and industrial automation. With its ability to
recognize multiple objects simultaneously, this project contributes to advancements in artificial
intelligence, making object detection more accessible, scalable, and practical for real-world
scenarios.

3.3 OBJECTIVE

Following are the major objectives of the proposed system:

● Develop a real-time object detection system for accurate and instant object recognition.
● Implement efficient and scalable processing to support multiple object categories.
● Enhance automation in security, traffic monitoring, and industrial applications.

GTU 5 NSIT
TeamID:736478 Internship Management
3.4 SCOPE(What it can do and can’t do)

The object detection system using OpenCV and MediaPipe is capable of real-time object
detection, tracking multiple objects simultaneously, and visualizing bounding boxes around
detected objects. It can estimate the distance of objects from the camera using polynomial
regression and identify user interactions, such as grabbing or moving an object. The system
includes a scoring mechanism based on these interactions and allows customization with
different pre-trained models for specific object categories. It also features efficient image
preprocessing techniques to optimize performance and provides real-time feedback by
dynamically displaying detected object names, distances, and user scores. Additionally, a user-
friendly interface can be integrated using Streamlit or OpenCV’s built-in display functions.

3.5 TECHNOLOGY AND LITERATURE REVIEW

This project utilizes computer vision and deep learning for real-time object detection. Python,
along with OpenCV for image processing and MediaPipe for object tracking, forms the core
technology stack. Supporting libraries like NumPy and cvzone aid in efficient computation and
visualization. The system operates through a webcam, enabling real-time interaction for
applications in gaming and monitoring.

Object detection has evolved from traditional feature extraction to AI-driven models like CNNs
and YOLO. Modern approaches, including MediaPipe, have improved accuracy, making real-
time detection more reliable for applications in security, automation, and augmented reality.

3.5.1 OpenCV

● OpenCV (Open Source Computer Vision Library) is an open-source library used for
computer vision and image processing tasks.
● OpenCV provides tools to analyze, modify, and interpret images and videos for
applications like object detection, face recognition, and motion tracking.
● OpenCV is optimized for real-time image processing, enabling fast and efficient object
detection from live video feeds.

GTU 6 NSIT
TeamID:736478 Internship Management
3.5.2 MediaPipe

● MediaPipe is an open-source framework designed to build machine learning (ML)

pipelines for real-time applications.
● It is optimized for low-latency and high-speed inference, making it ideal for applications
like gesture recognition, object tracking, and face detection.

3.5.3 Deep Learning Models

● Basic deep learning model that mimics the human brain using layers of neurons for
pattern recognition and classification.
● Specialized for image and video processing by using convolutional layers to extract
spatial features. Common in object detection, face recognition, and medical imaging.
● Designed for sequential data processing, such as speech recognition and language
modeling, by using memory through recurrent connections.
● It also allows us to continuously deliver our software by integrating with a large number
of testing and deployment technologies.

● An advanced type of RNN that overcomes vanishing gradient issues, making it effective
for long-term dependencies in time-series data, chatbots, and speech synthesis.

3.5.4 Distance Estimation Algorithm

● A computational method used to determine the distance between objects based on

mathematical models and sensor data.
● Uses signals, images, or reflections to measure the distance between an observer (sensor
or camera) and a target object.
● Distance Algorithm can rely on cameras, LiDAR, ultrasonic sensors, infrared, or radio
waves for estimation.

GTU 7 NSIT
TeamID:736478 Internship Management
3.5.5 Game Logic Engine

● A game logic engine in object detection is a computational framework that processes real-

time input from a camera to track objects (such as hands) and apply game mechanics

accordingly.

3.5.6 Graphical User Interface (GUI) Module

● A Graphical User Interface (GUI) module is a software component that enables users to
interact with an application through visually designed elements such as windows, buttons,
icons, and menus instead of text-based commands. It enhances user experience by making
software intuitive and interactive.

3.5.7 Data Processing Unit

● A Data Processing Unit (DPU) is a specialized processor designed to handle data-centric

tasks, offloading workloads from the CPU and GPU. It optimizes network, storage, and
security operations in high-performance computing environments. With no upfront costs
and a usage-based billing system, the cloud services described in this AWS overview
provide an affordable alternative to physical servers.

● The DPU processes data in parallel, efficiently managing data movement, encryption,
and acceleration. It reduces CPU overhead by handling networking, data compression,
and security functions.

3.6 PROJECT / INTERNSHIP PLANNING

Internship planning is part of project management, which relates to the use of schedules such as
charts to plan and subsequently report progress within the project environment. Initially, the
project scope is defined and the appropriate methods for completing the project are determined.
The durations for the various tasks necessary to complete the work are listed and grouped into a
work breakdown structure. Project planning is often used to organize different areas of a project,
including project plans, workloads and the management of teams and individuals.

GTU 8 NSIT
TeamID:736478 Internship Management
3.6.1 Project / Internship development approach and justification

A structured data science methodology ensures that the project follows a systematic and efficient
workflow. The key steps include:

● Faster and more accurate object detection.

● Reduced human effort in monitoring.
● Greater adaptability across various industries.
● More efficient real-time tracking systems.
● Better accuracy with continuous learning.
● Enhanced security and surveillance automation.

Defining Dataset Requirements – Identify the type, size, and quality of data needed for
accurate object detection. Consider factors like image resolution, object diversity, and
environmental conditions.

Collecting a Dataset – Gather relevant images or videos from various sources, ensuring
diversity for better model generalization. Data can be sourced from open datasets, real-world
captures, or synthetic generation.

Data Labelling – Annotate images with bounding boxes, segmentation masks, or keypoints to
define object positions. Proper labeling ensures the model learns object features accurately
during training.

GTU 9 NSIT
TeamID:736478 Internship Management
Choosing the Most Suitable ML Model – Select a deep learning architecture like YOLO, Faster
R-CNN, or SSD based on speed, accuracy, and hardware constraints. The model choice affects
performance and deployment feasibility.

Undertaking Several Training Iterations – Train the model using labeled data, adjusting
hyperparameters and loss functions. Multiple iterations improve accuracy by refining weights
and reducing detection errors.

Field Testing – Evaluate the trained model on real-world data to measure its accuracy and
robustness. Testing helps identify performance gaps and ensures adaptability to different
environments.

Integrating an ML Model into a Target Device – Deploy the optimized model on edge devices,
cloud systems, or mobile applications. This involves converting the model to a compatible
format and optimizing it for real-time inference.

3.6.2 Project / Internship Effort and Time, Cost Estimation

Estimation of Effort and Time

Activity Estimated Duration

Selecting project & supervisor 2 days

Project Initialization 1 week

Initial Proposal 2 weeks

Final Proposal 1 week and 3 days

Submission of Ethical Form 2 days

Building a Project Plan 1 week and 3 days

Presentation, Submission & Delivery 1 week and 5 days

Literature Review 2 days

Methodology 3 days

System Analysis 1 week and 4 days

Project Design 1 week

GTU 10 NSIT
TeamID:736478 Internship Management

Project Development 1 week

Testing 5 days

Report Writing 6 days

Total Time 13 weeks (around 100 days)

Project Cost Estimation

There are several factors which plays significant role in estimating the overall cost of project
development which are given below:

Hardware and Software Cost: It includes cost of various software tools which requires licenses
to operate as well as hardware running on the cloud with pay as you go facility

Testing and Quality Assurance Costs: This includes the cost of testing and quality assurance
activities, such as unit testing, integration testing, performance testing, and security testing, to
ensure the accuracy and reliability of development automation.

Documentation Costs: This includes the cost of training documenting the project deliverables,
including user manuals, technical documentation, and other project-related documentation.

Project Management Costs: This includes the cost of project management activities, such as
project planning, monitoring, and coordination, to ensure the smooth execution of the project and
timely delivery of project milestones.

3.7 INTERNSHIP SCHEDULING

● Project Scheduling is the culmination of a planning activity that is primary component of

software project management.
● When combined with estimation method sand risk analysis, scheduling, establishes a road
map for the project management.
● Scheduling begins with the process composition. The characteristics of the project are
used to adapt an appropriate task set for the work to be done.

GTU 11 NSIT
TeamID:736478
CHAPTER-4

SYSTEM ANALYSIS

4.1 STUDY OF CURRENT SYSTEM

● The existing system operates with a structured workflow but faces inefficiencies in data
processing, resource utilization, and real-time responsiveness. Users experience delays
due to outdated technology, manual interventions, and limited automation, affecting
overall productivity. Security vulnerabilities and compliance challenges further add to the
risks, making the system prone to data breaches. To enhance performance, scalability, and
user experience, a thorough system analysis is essential for identifying and addressing
these limitations.

4.2 PROBLEMS AND WEAKNESS OF CURRENT SYSTEM

● Lack of Efficiency – The system suffers from slow processing speeds and high manual
dependency, leading to delays in operations.
● Lack of Security Measures – Insufficient encryption, outdated authentication methods,
and weak access controls make the system vulnerable to cyber threats.
● Lack of Scalability – The current system cannot handle increased data loads or adapt to
growing user demands, limiting future expansion.

GTU 12 NSIT
TeamID:736478
4.3 SYSTEM FEASIBILITY

● Feasibility is the measure of how beneficial the development of information system will
be to an organization. Feasibility is the analysis of risks, costs and benefits relating to
economics, technology and user operation.
● There are several types of feasibility depending on the aspect they covers. Some
important feasibilities is as follows:
1. Operational Feasibility
2. Technical Feasibility
3. Scheduling Feasibility
4. Economic Feasibility

i. Operational Feasibility

● Operational feasibility is the measure of how well a proposed system solves the problems,
and takes advantage of the opportunities identified during scope definition and how it
satisfies the requirements identified in the requirements analysis phase of system
development.
● The system enhances real-time object detection, reducing manual intervention and
improving accuracy in tracking objects.
● It ensures smooth integration into existing workflows, making it user-friendly and
adaptable for various applications.

GTU 13 NSIT
TeamID:736478
ii. Technical Feasibility

● Utilizes advanced deep learning models like YOLO and Faster R-CNN, ensuring high
detection accuracy and real-time processing.

● The project is compatible with commonly used hardware like GPUs and edge devices,
making deployment feasible.

iii. Scheduling Feasibility

● The project follows a structured timeline, covering dataset collection, model training,
testing, and deployment within set deadlines.

● Development is feasible within the allocated time, ensuring timely delivery without
compromising system performance.

iv. Economic Feasibility

● The project leverages open-source tools like TensorFlow and OpenCV, minimizing costs
while maintaining high efficiency.
● Hardware requirements are optimized, reducing the need for expensive infrastructure
while ensuring smooth operation. Here are some ways that object detection can by
economically feasible:-
1. Using Open-Source Tools – Leveraging frameworks like TensorFlow, OpenCV, and
PyTorch reduces software licensing costs while maintaining high efficiency.

GTU 14 NSIT
TeamID:736478
2. Optimizing Hardware Utilization – Implementing lightweight models ensures that
the system runs efficiently on existing hardware, avoiding unnecessary upgrades.
3. Cloud-Based Deployment – Using cloud platforms like Google Colab or AWS for
training and inference eliminates the need for costly on-premise GPU setups.
4. Efficient Data Collection – Utilizing publicly available datasets or synthetic data
generation reduces expenses related to manual data collection and annotation.
5. Energy-Efficient Model Optimization – Using quantization and pruning techniques
lowers power consumption, reducing operational costs for real-time applications.
6. Scalable Implementation – Developing a modular system allows businesses to adopt
object detection gradually, minimizing upfront investment while enabling future
expansion.

GTU 15 NSIT
TeamID:736478

System Design

CHAPTER-5

SYSTEM DESIGN

5.1 Flowchart

GTU 16 NSIT
TeamID:736478
5.2 Use case Diagram

GTU 17 NSIT
TeamID:736478

CHAPTER-6

IMPLEMENTATION

6.1 Implementation of Developement Automation

● The object detection system is developed using OpenCV and MediaPipe, enabling real-
time recognition of hands and objects. A camera continuously captures frames, which are
processed for image enhancement and noise reduction.
● MediaPipe’s built-in models are used for detecting hands and gestures through predefined
tracking algorithms. These models efficiently track movements and key points, ensuring
accurate recognition.
● The system draws bounding boxes and key points on detected objects, highlighting their
position and movements. This visual representation helps in analyzing object interaction
in real time.
● The processed output is displayed on a graphical interface, providing continuous
feedback. The display dynamically updates as objects move, ensuring smooth
visualization for the user.
● Multi-threading is implemented to optimize processing speed, reducing lag and ensuring
smooth real-time performance. This allows the system to function efficiently even on
standard hardware.

GTU 18 NSIT
TeamID:736478
● Custom thresholding techniques are applied to improve object detection accuracy under
different lighting conditions. This enhances the system’s adaptability to varying
environments.
● The final implementation is tested for accuracy and performance in real-world scenarios.
Adjustments and optimizations are made to ensure reliable detection and tracking across
different use cases.

6.2 MODULES SPECIFICATION(S)

The object detection system consists of multiple modules working together for real-time tracking
and recognition. The image acquisition module captures frames from a webcam, while the
preprocessing module enhances image quality using OpenCV. MediaPipe’s object detection
module identifies hands, gestures, or objects and extracts key features for accurate recognition.

GTU 19 NSIT
TeamID:736478
6.3 SYSTEM INTERFACE

The system interface acts as a bridge between users and the object detection system, displaying
real-time detection results. It provides a visually interactive environment where detected objects,
bounding boxes, and key points are dynamically updated.
The following images display the project interface in operation, capturing real-time inputs and
processing them for object detection. These screenshots highlight how the system detects and
tracks objects, providing visual feedback with bounding boxes and key points.

GTU 20 NSIT
TeamID:736478

Conclusion and discussion

Chapter-7

CONCLUSION AND DISCUSSION

7.1 SUMMARY OF INTERNSHIP

In this internship I had worked on Devops methodology at Amar Infotech pvt.ltd, for this
internship I learnt so many best practices of Devops. I learnt about different technologies like
YAML, AWS, DOCKER, CHEF.

Firstly, I begin with the AWS cloud services because without a remote server I won't be able to
work on the project as EC2 was the only thing on which I deployed the application. Secondly,
read all the documentation of docker to create a container which will virtualize the server and
run the application on the lightweight environment which will not only increase the response
time but all utilize the server resources efficiently.

Probing Ahead, gain the knowledge about the ci cd pipelines and perform the development of
pipelines also configure smtp for the email notification which is invoked when an event occurs.

7.2DATES OF CONTINUOUS EVALUATION (CE-I & CE-II)

7.2.1 Review – I : 31/03/2022

GTU 21 NSIT
TeamID:736478
7.2.2 Review – II : 21/04/2022

7.2.2 Review – II : 28/04/2022

Conclusion and discussion

7.3 FUTURE ENHANCEMENT

In the future, I will continue to explore and develop the Devops CI CD pipeline, as well as its
integration into our other product development processes. Nevertheless, also try to automate
other pity parts of the development lifecycle.

7.4 REFERENCE

● Medium.com
● Stackoverflow.com
● Jenkins.io
● Hub.docker.com
● Stackexchange.com

GTU 22 NSIT