Occlusion-robust detector trained with occluded pedestrians

Zhixin Guo, Wenzhi Liao, Peter Veelaert, Wilfried Philips

Research output: Contribution to conferenceProceeding

1 Citation (Scopus)
5 Downloads (Pure)

Abstract

Pedestrian detection has achieved a remarkable progress in recent years, but challenges remain especially when occlusion happens. Intuitively, occluded pedestrian samples contain some characteristic occlusion appearance features that can help to improve detection. However, we have observed that most existing approaches intentionally avoid using samples of occluded pedestrians during the training stage. This is because such samples will introduce unreliable information, which affects the learning of model parameters and thus results in dramatic performance decline. In this paper, we propose a new framework for pedestrian detection. The proposed method exploits the use of occluded pedestrian samples to learn more robust features for discriminating pedestrians, and enables better performances on pedestrian detection, especially for the occluded pedestrians (which always happens in many real applications). Compared to some recent detectors on Caltech Pedestrian dataset, with our proposed method, detection miss rate for occluded pedestrians are significantly reduced.
Original languageEnglish
Pages86-94
Number of pages9
DOIs
Publication statusPublished - 18 Jan 2018
Event7th International Conference on Pattern Recognition Applications and Methods ICPRAM 2018 - Madeira, Portugal
Duration: 16 Jan 201818 Jan 2018

Conference

Conference7th International Conference on Pattern Recognition Applications and Methods ICPRAM 2018
Abbreviated titleICPRAM 2018
CountryPortugal
CityMadeira
Period16/01/1818/01/18

Keywords

  • pedestrian detection
  • occlusion handling
  • adaboost
  • integral channel features

Cite this

Guo, Z., Liao, W., Veelaert, P., & Philips, W. (2018). Occlusion-robust detector trained with occluded pedestrians. 86-94. 7th International Conference on Pattern Recognition Applications and Methods ICPRAM 2018, Madeira, Portugal. https://doi.org/10.5220/0006569200860094