Robust Automatic Motorcycle Helmet Violation Detection for an Intelligent Transportation System

Duong Nguyen-Ngoc Tran, Long Hoang Pham, Hyung-Joon Jeon, Huy-Hung Nguyen, Hyung-Min Jeon, Tai Huu-Phuong Tran, Jae Wook Jeon; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2023, pp. 5341-5349

Abstract

Video surveillance-based automatic detection of motorcycle helmet usage can enhance the effectiveness of educational and enforcement initiatives aimed at boosting road safety. Current detection methods, however, have room for enhancement, such as the inability to pinpoint individual motorcycles or differentiate between drivers and passengers in terms of helmet usage. This paper introduces a framework designed to detect and identify individual motorcycles while recording specific helmet usage for riders. The proposed classification approach for helmet usage demonstrates increased efficiency in comparison to previous research. Our findings highlight the exceptional accuracy of deep learning, with our method achieving a score of 0.7754 on the AI City 2023 Challenge Track 5 public leaderboard.

Related Material

[pdf]
[bibtex]
@InProceedings{Tran_2023_CVPR, author = {Tran, Duong Nguyen-Ngoc and Pham, Long Hoang and Jeon, Hyung-Joon and Nguyen, Huy-Hung and Jeon, Hyung-Min and Tran, Tai Huu-Phuong and Jeon, Jae Wook}, title = {Robust Automatic Motorcycle Helmet Violation Detection for an Intelligent Transportation System}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, month = {June}, year = {2023}, pages = {5341-5349} }