Minor PPT Yolo

You might also like

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 19

VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,

BACHUPALLY , HYDERABAD-500090
Department of Information Technology

OBJECT DETECTION AND TRACKING


SYSTEM USING DEEP LEARNING

Guide : M. Koteswara Rao

Project Team Members


D.Dinakar 20071A1274
K.Jathin 20071A1284
L.Ganesh 20071A1290

S. Anil Kumar 20071A12B1

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

AGENDA:
• Problem Statement
• Introduction
• Literature Survey
• Drawback Of Existing System
• Proposed System
• Methodology
• Objectives
• System Requirements
• conclusions
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

PROBLEM STATEMENT:

• Develop an efficient and accurate Helmetless bike riders


Detection using Deep Learning techniques to address the
growing demand for real-time and robust object detection
and tracking in various applications

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

INTRODUCTION:

• Nowadays most of the deaths in motorcycle accidents are


due to not wearing helmets.
• To detect the helmet less bike riders there are traditional
methods
• But they are often struggled with handling complex scenarios,
such as occlusions, varying object scales and cluttered
backgrounds.
• Recent advancements in deep learning have ushered in a
new era of object detection, providing remarkable accuracy
and adaptability in a wide array of settings.
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

LITERATURE SURVEY:

• CNN-Based Automatic Helmet Violation Detection of


Motorcyclists

• Real-Time Detection of Motorcyclist without Helmet using


Cascade of CNNs

• Detection of Helmetless Riders Using Faster R-CNN

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

1) CNN-Based Automatic Helmet Violation Detection of


Motorcyclists: (Toqeer Mahmood, Muhammad Asif)(2018)

● proposes a system to detect helmet violations automatically


from surveillance videos captured by roadside-mounted
cameras.
● proposed technique is based on faster region-based
convolutional neural network (R-CNN) deep learning model
that takes video as an input and performs helmet violation
detection to take necessary actions against traffic rule violators.
● Experimental analysis shows that the proposed system gives an
accuracy of 97.69%
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

2)Real-Time Detection of Motorcyclist without Helmet


using Cascade of CNNs (Dinesh Singh1, C. Vishnu)(2022)
Should be filled by dinakar
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

3) Detection of Helmetless Riders Using Faster R-CNN


, ,
(Meenu R,Sinta Raju Smrithi P Paul Swathy Sajeev)(2020)

Should be filled by jathin


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

DRAWBACKS OF EXISTING SYSTEM:


• Annotating a dataset for helmetless bike rider detection can be labor-
intensive and time-consuming
• CNN-based models, especially more complex architectures, may require
significant computational resources, which can affect their real-time
performance.
• CNN-based models, including R-CNN variants, can sometimes produce
false positives
• More complex CNN architectures like R-CNN can be computationally
expensive to train and deploy. They may require powerful GPUs or TPUs
for inference, limiting their use in resource-constrained environments.

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

PROPOSED SYSTEM:
• (YOLO) proposes using an end-to-end neural network that makes
predictions of bounding boxes and class probabilities all at once.
• It differs from the approach taken by previous object detection algorithms,
which repurposed classifiers to perform detection.
• YOLO achieved state-of-the-art results, beating other real-time object
detection algorithms by a large margin.
• YOLO performs all of its predictions with the help of a single fully
connected layer.
• Methods that use Region Proposal Networks perform multiple iterations for
the same image, while YOLO gets away with a single iteration

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

METHODOLOGY:

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

Architecture diagram of YOLOv3


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

ADVANTAGES:

1.YOLO can accurately identify and locate objects within an image, which is
crucial for detecting helmetless bike riders

2.performs object detection in a single pass through the neural network, making
it computationally efficient.

3.can be trained to detect a wide range of objects, including helmetless bike


riders.

4.It is robust when it comes to handling objects of different scales and


orientations in the same image.

5.It has open-source implementations and pretrained models available.

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

OUTPUTS:
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

OUTPUTS
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

Software and Hardware Requirements

Software Requirements
1.Python
Hardware Requirements
1.RAM-8gb or higher
2.Nvidia Geforce Graphics Card 4 gb or higher
3. System type-64 bit windows

Department of Information Technology


VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

CONCLUSION:

❖ It represents a significant step forward in enhancing road safety


and promoting responsible biking practices
❖ This project underscores the potential of AI-driven solutions to
address critical issues in our society.
❖ It not only showcases the effectiveness of YOLO in object
detection but also highlights the importance of leveraging
technology for the greater good.
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

REFERENCES:

● https://www.worldometers.info/w
orld-population/world-population-
by-year/
● https://www.hindawi.com/journals
/mpe/2022/8246776/
VNR VIGNANA JYOTHI INSTITUTE OF ENGINEERING AND TECHNOLOGY,
BACHUPALLY , HYDERABAD-500090
Department of Information Technology

You might also like