A Fast Traffic Sign Detection Algorithm Based on Three-Scale Nested Residual Structures

被引：0
作者：
Li X. [1 ,2 ]
Zhang J. [1 ,2 ]
Xie Z. [1 ,2 ]
Wang J. [1 ,2 ]
机构：
[1] School of Computer and Communication Engineering, Changsha University of Science and Technology, Changsha
[2] Hunan Provincial Key Laboratory of Intelligent Processing of Big Data on Transportation, Changsha University of Science and Technology, Changsha
来源：
Jisuanji Yanjiu yu Fazhan/Computer Research and Development | 2020年 / 57卷 / 05期
基金：
中国国家自然科学基金;
关键词：
Changsha University of Science and Technology (CSUST); CSUST Chinese traffic sign detection benchmark (CCTSDB); Multi-scale prediction; Nested residual network; Traffic sign detection; You only look once (YOLO) detection algorithm;
D O I：
10.7544/issn1000-1239.2020.20190445
中图分类号：
学科分类号：
摘要：
Automatic driving technology has high requirements for real-time and robustness of traffic sign detection in real world. The YOLOv3-tiny model is a lightweight network with good real-time performance in the object detection, but its accuracy is not high. In this paper, we use YOLOv3-tiny as the basic network and propose a fast traffic sign detection algorithm with three-scale nested residual structure. Firstly, shortcut based on pixel by pixel addition is employed in the YOLOv3-tiny network. It does not increase the number of feature map channels, and a small residual structure is formed in the network at the same time. Secondly, the predictive output with higher spatial resolution is also added through the shortcut, which contains more abundant spatial information, thus forming a large residual structure. Finally, the two residual structures are nested to form a three-scale predictive nested residual network, which makes the main network of Tiny located in these two residual structures and the parameters can be adjusted three times. The results show that the proposed algorithm can quickly and robustly detect traffic signs in real scenes. The F1 value of total traffic signs achieves 91.77% on German traffic sign detection benchmark and the detection time is 5ms. On CSUST Chinese traffic sign detection benchmark, F1 values of the Mandatory, the Prohibitory and the Warning are 92.41%, 93.91% and 92.03% respectively, and the detection time is 5ms. © 2020, Science Press. All right reserved.
引用
收藏
页码：1022 / 1036
页数：14
相关论文
共 26 条
  • [1] Le T.T., Tran S.T., Mita S., Et al., Real time traffic sign detection using color and shape-based features, Proc of the 2nd Asian Conf on Intelligent Information and Database Systems, pp. 268-278, (2010)
  • [2] Zhang T., Zou J., Jia W., Fast and robust road sign detection in driver assistance systems, Applied Intelligence, 48, 11, pp. 4113-4127, (2018)
  • [3] Gu M., Cai Z., Traffic sign recognition based on parameter-free detector and DT-CWT, Journal of Computer Research and Development, 50, 9, pp. 1893-1901, (2013)
  • [4] Xu X., Jin J., Zhang S., Et al., Smart data driven traffic sign detection method based on adaptive color threshold and shape symmetry, Future Generation Computer Systems, 94, pp. 381-391, (2019)
  • [5] Zhang J., He Y., Dai Y., Et al., Multi-feature fusion based circular traffic sign detection, Pattern Recognition and Artificial Intelligence, 24, 2, pp. 226-232, (2011)
  • [6] Girshick R., Donahue J., Darrell T., Et al., Rich feature hierarchies for accurate object detection and semantic segmentation, Proc of the 27th IEEE Conf on Computer Vision and Pattern Recognition, pp. 580-587, (2014)
  • [7] He K., Zhang X., Ren S., Et al., Spatial pyramid pooling in deep convolutional networks for visual recognition, IEEE Transactions on Pattern Analysis and Machine Intelligence, 37, 9, pp. 1904-1916, (2015)
  • [8] Girshick R., Fast R-CNN, Proc of the 15th IEEE Int Conf on Computer Vision, pp. 1440-1448, (2015)
  • [9] Ren S., He K., Girshick R., Et al., Faster R-CNN: Towards real-time object detection with region proposal networks, Proc of Advances in Neural Information Processing Systems, pp. 91-99, (2015)
  • [10] Yang T., Long X., Sangaiah A.K., Et al., Deep detection network for real-life traffic sign in vehicular networks, Computer Networks, 136, pp. 95-104, (2018)
123