An Occlusion-Robust Feature Selection Framework in Pedestrian Detection †
Better features have been driving the progress of pedestrian detection over the past years. However, as features become richer and higher dimensional, noise and redundancy in the feature sets become bigger problems. These problems slow down learning and can even reduce the performance of the learned...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2018-07-01
|
Series: | Sensors |
Subjects: | |
Online Access: | http://www.mdpi.com/1424-8220/18/7/2272 |
id |
doaj-133cdd148c8f4c5cb88f3450c04059d2 |
---|---|
record_format |
Article |
spelling |
doaj-133cdd148c8f4c5cb88f3450c04059d22020-11-24T22:00:06ZengMDPI AGSensors1424-82202018-07-01187227210.3390/s18072272s18072272An Occlusion-Robust Feature Selection Framework in Pedestrian Detection †Zhixin Guo0Wenzhi Liao1Yifan Xiao2Peter Veelaert3Wilfried Philips4Department of Telecommunications and Information Processing, Ghent University-Interuniversitair Micro-Elektronica Centrum (IMEC), Sint-Pietersnieuwstraat 41, 9000 Gent, BelgiumDepartment of Telecommunications and Information Processing, Ghent University-Interuniversitair Micro-Elektronica Centrum (IMEC), Sint-Pietersnieuwstraat 41, 9000 Gent, BelgiumSchool of Information Science and Engineering, Shandong University, Jinan 250100, ChinaDepartment of Telecommunications and Information Processing, Ghent University-Interuniversitair Micro-Elektronica Centrum (IMEC), Sint-Pietersnieuwstraat 41, 9000 Gent, BelgiumDepartment of Telecommunications and Information Processing, Ghent University-Interuniversitair Micro-Elektronica Centrum (IMEC), Sint-Pietersnieuwstraat 41, 9000 Gent, BelgiumBetter features have been driving the progress of pedestrian detection over the past years. However, as features become richer and higher dimensional, noise and redundancy in the feature sets become bigger problems. These problems slow down learning and can even reduce the performance of the learned model. Current solutions typically exploit dimension reduction techniques. In this paper, we propose a simple but effective feature selection framework for pedestrian detection. Moreover, we introduce occluded pedestrian samples into the training process and combine it with a new feature selection criterion, which enables improved performances for occlusion handling problems. Experimental results on the Caltech Pedestrian dataset demonstrate the efficiency of our method over the state-of-art methods, especially for the occluded pedestrians.http://www.mdpi.com/1424-8220/18/7/2272pedestrian detectionfeature selectionocclusion handlingdeep learning |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Zhixin Guo Wenzhi Liao Yifan Xiao Peter Veelaert Wilfried Philips |
spellingShingle |
Zhixin Guo Wenzhi Liao Yifan Xiao Peter Veelaert Wilfried Philips An Occlusion-Robust Feature Selection Framework in Pedestrian Detection † Sensors pedestrian detection feature selection occlusion handling deep learning |
author_facet |
Zhixin Guo Wenzhi Liao Yifan Xiao Peter Veelaert Wilfried Philips |
author_sort |
Zhixin Guo |
title |
An Occlusion-Robust Feature Selection Framework in Pedestrian Detection † |
title_short |
An Occlusion-Robust Feature Selection Framework in Pedestrian Detection † |
title_full |
An Occlusion-Robust Feature Selection Framework in Pedestrian Detection † |
title_fullStr |
An Occlusion-Robust Feature Selection Framework in Pedestrian Detection † |
title_full_unstemmed |
An Occlusion-Robust Feature Selection Framework in Pedestrian Detection † |
title_sort |
occlusion-robust feature selection framework in pedestrian detection † |
publisher |
MDPI AG |
series |
Sensors |
issn |
1424-8220 |
publishDate |
2018-07-01 |
description |
Better features have been driving the progress of pedestrian detection over the past years. However, as features become richer and higher dimensional, noise and redundancy in the feature sets become bigger problems. These problems slow down learning and can even reduce the performance of the learned model. Current solutions typically exploit dimension reduction techniques. In this paper, we propose a simple but effective feature selection framework for pedestrian detection. Moreover, we introduce occluded pedestrian samples into the training process and combine it with a new feature selection criterion, which enables improved performances for occlusion handling problems. Experimental results on the Caltech Pedestrian dataset demonstrate the efficiency of our method over the state-of-art methods, especially for the occluded pedestrians. |
topic |
pedestrian detection feature selection occlusion handling deep learning |
url |
http://www.mdpi.com/1424-8220/18/7/2272 |
work_keys_str_mv |
AT zhixinguo anocclusionrobustfeatureselectionframeworkinpedestriandetection AT wenzhiliao anocclusionrobustfeatureselectionframeworkinpedestriandetection AT yifanxiao anocclusionrobustfeatureselectionframeworkinpedestriandetection AT peterveelaert anocclusionrobustfeatureselectionframeworkinpedestriandetection AT wilfriedphilips anocclusionrobustfeatureselectionframeworkinpedestriandetection AT zhixinguo occlusionrobustfeatureselectionframeworkinpedestriandetection AT wenzhiliao occlusionrobustfeatureselectionframeworkinpedestriandetection AT yifanxiao occlusionrobustfeatureselectionframeworkinpedestriandetection AT peterveelaert occlusionrobustfeatureselectionframeworkinpedestriandetection AT wilfriedphilips occlusionrobustfeatureselectionframeworkinpedestriandetection |
_version_ |
1725845339125055488 |