Convolutional Neural Networks for Traffic Sign Recognition

Right-of-way image-based traffic sign recognition (TSR) is an important research field in intelligent transportation systems. Convolutional neural networks (CNNs) have made breakthroughs in TSR in recent years. However, the traditional convolution lacks invariance for affine transformations such as translation, scaling, shearing and rotation of symbols. To preserve spatial invariance of traffic signs, a Spatial Transformer-Convolutional Neural Network (ST-CNN) is proposed in this paper, where a Spatial Transformer Network (STN) is placed in front of different convolution modules. This method transforms images that are not easily segmented in the original image spatial into the feature spatial which is centered on the reference image and realizes the classification function. This paper uses the German Traffic Sign Recognition Benchmark (GTSRB) dataset for training and testing. The performance of different STNs in the main network location is analyzed and the best model is selected. The accuracy in GTSRB is 99.36%.

• Record URL:
  • https://doi.org/10.1061/9780784483565.039
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/isbn/9780784483565
• Supplemental Notes:
  • © 2021 American Society of Civil Engineers.
• Corporate Authors:

  American Society of Civil Engineers

  1801 Alexander Bell Drive
  Reston, VA  United States  20191-4400
• Authors:
  • Wei, Zhonghua
  • Gu, Heng
  • Zhang, Ran
  • Peng, Jingxuan
  • Qui, Shi
• Conference:
  • 21st COTA International Conference of Transportation Professionals
  • Location: Xi'an , China
  • Date: 2021-12-16 to 2021-12-19
• Publication Date: 2021

Language

• English

Media Info

• Media Type: Web
• Pagination: pp 399-409
• Monograph Title: CICTP 2021: Advanced Transportation, Enhanced Connection

Subject/Index Terms

• TRT Terms: Detection and identification systems; Intelligent transportation systems; Neural networks; Right of way (Traffic); Traffic signs
• Subject Areas: Highways; Operations and Traffic Management;

Filing Info

• Accession Number: 01831715
• Record Type: Publication
• ISBN: 9780784483565
• Files: TRIS, ASCE
• Created Date: Dec 28 2021 9:32AM