Improving the Performance of Universal Traffic Light Recognition Through Dataset Construction and Selection in Semi-supervised Learning

被引：0

作者：

Kim, Dayoung ^{[1
]}

Li, Xingyou ^{[1
]}

Kim, Hakil ^{[1
]}

机构：

[1] Department of Electrical and Computer Engineering, Inha University

来源：

Journal of Institute of Control, Robotics and Systems | 2024年 / 30卷 / 08期

关键词：

Data selection; Semi-Supervised Object Detection (SSOD); Traffic light recognition;

D O I：

10.5302/J.ICROS.2024.24.0120

中图分类号：

学科分类号：

摘要：

Traffic light recognition in autonomous driving is an essential but very challenging task because its performance is affected by unpredictable environmental conditions. Moreover, the shapes and installations of traffic lights in various countries require a lot of cost and time to collect a huge dataset and train a deep learning model. This study develops a method for constructing training datasets with minimal resource expenditure for more efficient and universally applicable traffic light recognition. It aims to develop a deep learning scheme that can universally recognize various international traffic light configurations by replacing the initial supervised learning stage with a phase that can recognize multiple types of traffic lights and improve performance through self-supervised learning on new datasets. Additionally, this study introduces a data selection algorithm that enables robust recognition under changes in road conditions by enhancing adaptability across diverse environments. This research not only demonstrates improved recognition performance in hazardous road situations but also concludes with potential for efficient application worldwide. Experimental results showed 24.2% improvement in mAP50, even with the same proportion of labeled data, and similar performance with half the labeled data. © ICROS 2024.

引用

页码：787 / 792

页数：5

共 23 条

[1]

Machardy Z., Khan A., Obana K., Iwashina S., V2X access technologies: Regulation, research, and remaining challenges, IEEE Communications Surveys & Tutorials, 20, 3, pp. 1858-1877, (2018)

[2]

Weber M., Wolf P., Zollner J.M., DeepTLR: A single deep convolutional network for detection and classification of traffic lights, 2016 IEEE Intelligent Vehicles Symposium (IV), IEEE, pp. 342-348, (2016)

[3]

Hwang Y., Lee Y., Guest Editorial: Industrial applications of deep learning technology, Journal of Institute of Control, 29, 12, (2023)

[4]

Ouyang Z., Niu J., Liu Y., Guizani M., Deep CNN-based real-time traffic light detector for self-driving vehicles, IEEE Transactions on Mobile Computing, 19, 2, pp. 300-313, (2019)

[5]

Lee J., Baek S., Choi E., Moon H., Recognition of real-time traffic signals using the algorithm fusion of on-board graphic chip-based machine learning and deep learning, Journal of Institute of Control, 28, 2, pp. 88-94, (2022)

[6]

Felder R.M., Brent R., Active learning: An introduction, ASQ Higher Education Brief, 2, 4, pp. 1-5, (2009)

[7]

Sohn K., Zhang Z., Li C.L., Zhang H., Lee C.Y., Pfister T., A Simple Semi-Supervised Learning Framework for Object Detection, (2020)

[8]

Xu M., Zhang Z., Hu H., Wang J., Wang L., Wei F., Liu Z., End-to-end semi-supervised object detection with soft teacher, Proceedings of the IEEE/CVF International Conference on Computer Vision, (2021)

[9]

Siva P., Russell C., Xiang T., Agapito L., Looking beyond the image: Unsupervised learning for object saliency and detection, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, (2013)

[10]

Buchert F., Navab N., Kim S., Exploiting diversity of unlabeled data for label-efficient semi-supervised active learning, 2022 26Th International Conference on Pattern Recognition (ICPR), IEEE, (2022)

← 1 2 3 →