How Far are We from Solving Pedestrian Detection?

被引:194
作者
Zhang, Shanshan [1 ]
Benenson, Rodrigo [1 ]
Omran, Mohamed [1 ]
Hosang, Jan [1 ]
Schiele, Bernt [1 ]
机构
[1] Max Planck Inst Informat, Saarbrucken, Germany
来源
2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR) | 2016年
关键词
D O I
10.1109/CVPR.2016.141
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Encouraged by the recent progress in pedestrian detection, we investigate the gap between current state-of-the-art methods and the "perfect single frame detector". We enable our analysis by creating a human baseline for pedestrian detection (over the Caltech dataset), and by manually clustering the recurrent errors of a top detector. Our results characterise both localisation and background-versus-foreground errors. To address localisation errors we study the impact of training annotation noise on the detector performance, and show that we can improve even with a small portion of sanitised training data. To address background/ foreground discrimination, we study convnets for pedestrian detection, and discuss which factors affect their performance. Other than our in-depth analysis, we report top performance on the Caltech dataset, and provide a new sanitised set of training and test annotations.
引用
收藏
页码:1259 / 1267
页数:9
相关论文
共 23 条
[1]  
Agrawal P, 2014, LECT NOTES COMPUT SC, V8695, P329, DOI 10.1007/978-3-319-10584-0_22
[2]  
[Anonymous], 2014, P 27 INT C NEURAL IN
[3]  
[Anonymous], 2015, ABSTR APPL AN, DOI DOI 10.1016/J.JNCA.2015.03.007
[4]  
[Anonymous], PAMI
[5]  
[Anonymous], DAGM
[6]   Ten Years of Pedestrian Detection, What Have We Learned? [J].
Benenson, Rodrigo ;
Omran, Mohamed ;
Hosang, Jan ;
Schiele, Bernt .
COMPUTER VISION - ECCV 2014 WORKSHOPS, PT II, 2015, 8926 :613-627
[7]   Seeking the strongest rigid detector [J].
Benenson, Rodrigo ;
Mathias, Markus ;
Tuytelaars, Tinne ;
Van Gool, Luc .
2013 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2013, :3666-3673
[8]   Histograms of oriented gradients for human detection [J].
Dalal, N ;
Triggs, B .
2005 IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, VOL 1, PROCEEDINGS, 2005, :886-893
[9]   Fast Feature Pyramids for Object Detection [J].
Dollar, Piotr ;
Appel, Ron ;
Belongie, Serge ;
Perona, Pietro .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2014, 36 (08) :1532-1545
[10]   Pedestrian Detection: An Evaluation of the State of the Art [J].
Dollar, Piotr ;
Wojek, Christian ;
Schiele, Bernt ;
Perona, Pietro .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2012, 34 (04) :743-761