Cover Image

Vision Based Pedestrian Traffic Counting Using Deep Learning Method

Yolov4 is a one stage detector to the vision-based object detection system. It is a predictive technique that provides faster and accurate results with minimal background errors. Object detection is a computer vision technique that performs to identify and locate objects within an image or video. In...

Full description

Saved in:
Bibliographic Details
Main Author: Ho, Chia Hui
Format: Undergraduates Project Papers
Language:English
Published: 2022
Subjects:
TA Engineering (General). Civil engineering (General)
TK Electrical engineering. Electronics Nuclear engineering
Online Access:http://umpir.ump.edu.my/id/eprint/39914/1/EA18177_Ho_Thesis%20-%20Ho%20Chia%20Hui.pdf
http://umpir.ump.edu.my/id/eprint/39914/
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.ump.umpir.39914
record_format eprints
spelling my.ump.umpir.399142024-01-08T10:36:03Z http://umpir.ump.edu.my/id/eprint/39914/ Vision Based Pedestrian Traffic Counting Using Deep Learning Method Ho, Chia Hui TA Engineering (General). Civil engineering (General) TK Electrical engineering. Electronics Nuclear engineering Yolov4 is a one stage detector to the vision-based object detection system. It is a predictive technique that provides faster and accurate results with minimal background errors. Object detection is a computer vision technique that performs to identify and locate objects within an image or video. In other word, object detection draws bounding boxes around these detected objects, which allow us to know where objects are in. One of the challenges of object detection is occlusion reduce the detection accuracy. The aim of this project is to detect and track the pedestrian even they are walk in group. The output of the bounding box is obtained after the input image passed through the Yolov4 network architecture. After that the threshold and non-maximum suppression (NMS) are applied to get the best bounding box. The counting function is done when after NMS. Score threshold is adjustable to observe which thresholds can get a better accuracy result in an image or video. The accuracy is obtained by applying the formula of TP, TN, FP and FN. The result shown that using score threshold of 0.35 can get higher accuracy which is 84.62%~100% after simulate. 2022-06 Undergraduates Project Papers NonPeerReviewed pdf en http://umpir.ump.edu.my/id/eprint/39914/1/EA18177_Ho_Thesis%20-%20Ho%20Chia%20Hui.pdf Ho, Chia Hui (2022) Vision Based Pedestrian Traffic Counting Using Deep Learning Method. College of Engineering, Universiti Malaysia Pahang Al-Sultan Abdullah.
institution Universiti Malaysia Pahang Al-Sultan Abdullah
building UMPSA Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Malaysia Pahang Al-Sultan Abdullah
content_source UMPSA Institutional Repository
url_provider http://umpir.ump.edu.my/
language English
topic TA Engineering (General). Civil engineering (General)
TK Electrical engineering. Electronics Nuclear engineering
spellingShingle TA Engineering (General). Civil engineering (General)
TK Electrical engineering. Electronics Nuclear engineering
Ho, Chia Hui
Vision Based Pedestrian Traffic Counting Using Deep Learning Method
description Yolov4 is a one stage detector to the vision-based object detection system. It is a predictive technique that provides faster and accurate results with minimal background errors. Object detection is a computer vision technique that performs to identify and locate objects within an image or video. In other word, object detection draws bounding boxes around these detected objects, which allow us to know where objects are in. One of the challenges of object detection is occlusion reduce the detection accuracy. The aim of this project is to detect and track the pedestrian even they are walk in group. The output of the bounding box is obtained after the input image passed through the Yolov4 network architecture. After that the threshold and non-maximum suppression (NMS) are applied to get the best bounding box. The counting function is done when after NMS. Score threshold is adjustable to observe which thresholds can get a better accuracy result in an image or video. The accuracy is obtained by applying the formula of TP, TN, FP and FN. The result shown that using score threshold of 0.35 can get higher accuracy which is 84.62%~100% after simulate.
format Undergraduates Project Papers
author Ho, Chia Hui
author_facet Ho, Chia Hui
author_sort Ho, Chia Hui
title Vision Based Pedestrian Traffic Counting Using Deep Learning Method
title_short Vision Based Pedestrian Traffic Counting Using Deep Learning Method
title_full Vision Based Pedestrian Traffic Counting Using Deep Learning Method
title_fullStr Vision Based Pedestrian Traffic Counting Using Deep Learning Method
title_full_unstemmed Vision Based Pedestrian Traffic Counting Using Deep Learning Method
title_sort vision based pedestrian traffic counting using deep learning method
publishDate 2022
url http://umpir.ump.edu.my/id/eprint/39914/1/EA18177_Ho_Thesis%20-%20Ho%20Chia%20Hui.pdf
http://umpir.ump.edu.my/id/eprint/39914/
_version_ 1822924043932663808
score 13.232414

Similar Items