Convolution Pedestrian Detection Based on Random Fusion of Color and Gradient
- DOI
- 10.2991/wartia-17.2017.6How to use a DOI?
- Keywords
- double color; channel switching; extended operator; random fusion;
- Abstract
The complex prospects in pedestrian detection, such as backpacks and other obstacles, are likely to cause interference to pedestrians. Since previous pedestrian detection can only use separate gradient information, the color information is neglected, and the gradient direction information is not accurate because of noise. In this paper, we propose a convolution network based on the combination of double color and improved Sobel extended gradient information to detect pedestrians and other prospects. The model combines convolution of RGB and HSI color channels and improved Sobel extended gradient fusion channels respectively. Then the stochastic fusion feature vector method is proposed to fuse the color and gradient information randomly, and the final result of pedestrian detection is obtained. Experimental results show that the proposed method improves the detection accuracy.
- Copyright
- © 2017, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Xiangquan Gui AU - Jiajun Jiang AU - Li Li AU - Dongmei Chen AU - Lei Gao PY - 2017/11 DA - 2017/11 TI - Convolution Pedestrian Detection Based on Random Fusion of Color and Gradient BT - Proceedings of the 3rd Workshop on Advanced Research and Technology in Industry (WARTIA 2017) PB - Atlantis Press SP - 26 EP - 34 SN - 2352-5401 UR - https://doi.org/10.2991/wartia-17.2017.6 DO - 10.2991/wartia-17.2017.6 ID - Gui2017/11 ER -