Repulsion Loss: Detecting Pedestrians in a Crowd
Abstract
Detecting individual pedestrians in a crowd remains a challenging problem since the pedestrians often gather together and occlude each other in real-world scenarios. In this paper, we first explore how a state-of-the-art pedestrian detector is harmed by crowd occlusion via experimentation, providing insights into the crowd occlusion problem. Then, we propose a novel bounding box regression loss specifically designed for crowd scenes, termed repulsion loss. This loss is driven by two motivations: the attraction by target, and the repulsion by other surrounding objects. The repulsion term prevents the proposal from shifting to surrounding objects thus leading to more crowd-robust localization. Our detector trained by repulsion loss outperforms all the state-of-the-art methods with a significant improvement in occlusion cases.
- Publication:
-
arXiv e-prints
- Pub Date:
- November 2017
- DOI:
- 10.48550/arXiv.1711.07752
- arXiv:
- arXiv:1711.07752
- Bibcode:
- 2017arXiv171107752W
- Keywords:
-
- Computer Science - Computer Vision and Pattern Recognition
- E-Print:
- Accepted to IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2018