基于改进YOLOv4-Tiny算法的移动端实时司机违章行为检测
Real-Time Drivers’ Violation Detection on Mobile Terminal Based on Improved YOLOv4-Tiny
摘要: 司机违章行为在日常生活中屡见不鲜，为解决在移动端上高精度、实时监测司机的驾驶行为的问题，基于一种轻量级的目标检测算法YOLOv4-tiny，通过引入跨通道部分的连接机制，减少了模型的参数，在YOLOv3-tiny的基础上改进残差的思想，并在损失函数上将定位损失替换为CIOU损失实现更精确的边框，最后通过知识蒸馏的手段，利用教师模型YOLOv4指导tiny训练进一步提升其性能。通过在公开数据集和自建数据集上的多项实验对比结果显示，YOLOv4-tiny达到了更高的精度(分别提升7%和10%)，实现了对算力较低的嵌入式设备或移动端的实时检测。
Abstract: Driver violations are common in daily life. In order to solve the problem of monitoring drivers’ driving behavior on mobile with high accuracy and in real time, based on a lightweight target detection algorithm YOLOv4-tiny, the parameters of the model are reduced by introducing a connection mechanism across channel sections, improving the idea of residuals based on YOLOv3-tiny, and replacing the localization loss with CIOU loss in the loss function to achieve more accurate edges, and finally, by means of knowledge distillation, the teacher model YOLOv4 is used to guide tiny training to further improve its performance. The results of several experiments comparing public and self-built datasets show that YOLOv4-tiny achieves higher accuracy (7% and 10% improvement, respectively) and realizes real-time detection on embedded devices or mobile with low computing power.
文章引用：秦丹峰, 尹相辉, 龚学余. 基于改进YOLOv4-Tiny算法的移动端实时司机违章行为检测[J]. 计算机科学与应用, 2021, 11(5): 1291-1300. https://doi.org/10.12677/CSA.2021.115131

参考文献

[1] Ren, S., He, K., Girshick, R., et al. (2017) Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39, 1137-1149.
[2] Redmon, J., Divvala, S., Girshick, R., et al. (2016) You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, 27-30 June 2016, 779-788. [Google Scholar] [CrossRef] 
[3] Redmon, J. and Farhadi, A. (2017) YOLO9000: Better, Faster, Stronger. /Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, 21-26 July 2017, 7263-7271. [Google Scholar] [CrossRef] 
[4] Redmon, J. and Farhadi, A. (2018) YOLOv3: An Incremental Im-provement. arXiv:1804.02767 [cs.CV]
[5] Bochkovskiy, A., Wang, C.Y. and Liao, H.Y.M. (2020) YOLOv4: Opti-mal Speed and Accuracy of Object Detection. arXiv:2004.10934 [cs.CV]
[6] He, K., Zhang, X., Ren, S., et al. (2016) Deep Residual Learning for Image Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, 27-30 June 2016, 770-778. [Google Scholar] [CrossRef] 
[7] Wang, C.Y., Liao, H.Y.M., Wu, Y.H., et al. (2020) CSPNet: A New Backbone That Can Enhance Learning Capability of CNN. Proceedings of the IEEE/CVF Conference on Computer Vi-sion and Pattern Recognition Workshops, Seattle, 14-19 June 2020, 390-391. [Google Scholar] [CrossRef] 
[8] He, K., Zhang, X., Ren, S., et al. (2015) Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 37, 1904-1916. [Google Scholar] [CrossRef] 
[9] Zheng, Z., Wang, P., Liu, W., et al. (2020) Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression. Proceedings of the AAAI Conference on Artificial Intelligence, 34, AAAI-20 Technical Tracks 7. [Google Scholar] [CrossRef] 
[10] Yun, S., Han, D., Oh, S.J., et al. (2019) CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features. Proceedings of the IEEE International Conference on Computer Vision, Seoul, 27 October-2 November 2019, 6023-6032. [Google Scholar] [CrossRef] 
[11] Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K. and Li, F.-F. (2009) ImageNet: A Large-Scale Hierarchical Image Database. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Miami, 20-25 June 2009, 248-255. [Google Scholar] [CrossRef]