Enhancing public transportation detection using YOLOv5

Arbi Shukhair, Nur Liyana Ameera and Mansor, Hasmah and Gunawan, Teddy Surya and Abdul Kadir, Kushsairy (2024) Enhancing public transportation detection using YOLOv5. In: IEEE 9th International Conference on Smart Instrumentation, Measurement and Applications (ICSIMA 2023), 17th - 18th October 2023, Kuala Lumpur, Malaysia.

PDF (Full Paper) - Published Version Restricted to Registered users only Download (1MB) | Request a copy

Abstract

In the current context of urbanization and transportation expansion, the need for accurate and efficient detection systems for public transportation has become of the utmost importance. The paper presents a novel strategy to establish new standards in transportation detection systems. Using the power of the YOLOv5 deep learning algorithm, the dataset is divided into training, testing, and validation segments to ensure a thorough evaluation. With a training dataset size of 75% and a test-validation split of 25%, our methodology showcases a compelling mean Average Precision (mAP) value of 0.973. Our findings highlight a precision of 0.971, pointing to accurate predictions in approximately 97.1% of cases, and a recall of 0.953, underscoring the model's efficiency in capturing around 95.3% of relevant objects. Such results, particularly the distinguishable taxi class among similar objects, represent significant improvements over previous benchmarks. The model's prowess is evident in its ability to distinguish even in situations involving entities that closely resemble one another, such as taxis and police cars. Our proposed system excels with increased accuracy, precision, and F1 scores compared to a standard study. This paper concludes that with the strategic application of YOLOv5, the future of public transportation detection systems is bright and on the cusp of a new era of efficiency and precision.

Item Type:	Proceeding Paper (Invited Papers)
Additional Information:	First author is student. So, Teddy Gunawan is the corresponding author from IIUM. External collaboration with UniKL BMI (national).
Uncontrolled Keywords:	public transportation, object detection, computer vision, deep learning, YOLOv5
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7800 Electronics. Computer engineering. Computer hardware. Photoelectronic devices > TK7885 Computer engineering
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Engineering Kulliyyah of Engineering > Department of Electrical and Computer Engineering
Depositing User:	Prof. Dr. Teddy Surya Gunawan
Date Deposited:	15 Jan 2024 09:47
Last Modified:	23 Feb 2024 15:43
URI:	http://irep.iium.edu.my/id/eprint/110155

Actions (login required)

View Item