Ghent University Academic Bibliography

Advanced

Occlusion-robust Detector Trained with Occluded Pedestrians

Zhixin Guo, Wenzhi Liao, Peter Veelaert and Wilfried Philips (2018)
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.
Please use this url to cite or link to this publication:
author
organization
year
type
conference
publication status
published
keyword
Pedestrian Detection, Occlusion Handling, Adaboost, Integral Channel Features
publisher
SCITEPRESS - Science and Technology Publications
ISBN
9789897582769
DOI
10.5220/0006569200860094
UGent publication?
yes
classification
U
id
8569163
handle
http://hdl.handle.net/1854/LU-8569163
date created
2018-07-13 15:20:36
date last changed
2018-07-13 15:20:45
@inproceedings{8569163,
  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.},
  author       = {Guo, Zhixin and Liao, Wenzhi and Veelaert, Peter and Philips, Wilfried},
  isbn         = {9789897582769},
  keyword      = {Pedestrian Detection,Occlusion Handling,Adaboost,Integral Channel Features},
  publisher    = {SCITEPRESS - Science and Technology Publications},
  title        = {Occlusion-robust Detector Trained with Occluded Pedestrians},
  url          = {http://dx.doi.org/10.5220/0006569200860094},
  year         = {2018},
}

Chicago
Guo, Zhixin, Wenzhi Liao, Peter Veelaert, and Wilfried Philips. 2018. “Occlusion-robust Detector Trained with Occluded Pedestrians.” In SCITEPRESS - Science and Technology Publications.
APA
Guo, Zhixin, Liao, W., Veelaert, P., & Philips, W. (2018). Occlusion-robust Detector Trained with Occluded Pedestrians. SCITEPRESS - Science and Technology Publications.
Vancouver
1.
Guo Z, Liao W, Veelaert P, Philips W. Occlusion-robust Detector Trained with Occluded Pedestrians. SCITEPRESS - Science and Technology Publications; 2018.
MLA
Guo, Zhixin, Wenzhi Liao, Peter Veelaert, et al. “Occlusion-robust Detector Trained with Occluded Pedestrians.” SCITEPRESS - Science and Technology Publications, 2018. Print.