Faster R-CNN for small traffic sign detection

Zhuo Zhang, Xiaolong Zhou, Sixian Chan, Shengyong Chen, Honghai Liu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1233 Downloads (Pure)


Traffic sign detection is essential in autonomous driving. It is challenging especially when large proportion of instance to be detected are in small size. Directly applying state-of-the-art object detection algorithm Faster R-CNN for small traffic sign detection renders unsatisfactory detection rate, while a higher accuracy will be performed if the input images are upsampled. In this paper, we first investigate Faster R-CNN’s network architecture, and regard its weak performance on small instances as improper receptive field. Then we augment its architecture according to receptive field with a higher accuracy achieved and no obvious incremental computational cost. Experiments are conducted to validate the effectiveness of proposed method and give an comparison to the state-of-the-art detection algorithms on both accuracy and computational cost. The experimental results demonstrate an improved detection accuracy and an competitive computing speed of the proposed method.
Original languageEnglish
Title of host publicationComputer Vision - 2nd CCF Chinese Conference, CCCV 2017, Proceedings
Subtitle of host publicationSecond CCF Chinese Conference, CCCV 2017, Tianjin, China, October 11–14, 2017, Proceedings, Part III
EditorsJinfeng Yang, Qingshan Liu, Liang Wang, Xiang Bai, Qinghua Hu, Ming-Ming Cheng, Deyu Meng
PublisherSpringer Verlag
Number of pages11
ISBN (Electronic)978-981-10-7305-2
ISBN (Print)978-981-10-7304-5
Publication statusPublished - Dec 2017
EventCCF Chinese Conference on Computer Vision - Tianjin, China
Duration: 11 Oct 201714 Oct 2017

Publication series

NameCommunications in Computer and Information Science
ISSN (Print)1865-0929


ConferenceCCF Chinese Conference on Computer Vision
Abbreviated titleCCCV 2017


Dive into the research topics of 'Faster R-CNN for small traffic sign detection'. Together they form a unique fingerprint.

Cite this