TY - JOUR
T1 - Robustness Analysis of Pedestrian Detectors for Surveillance
AU - Fang, Yuming
AU - Ding, Guanqun
AU - Yuan, Yuan
AU - Lin, Weisi
AU - Liu, Haiwen
N1 - Publisher Copyright:
© 2013 IEEE.
PY - 2018/5/23
Y1 - 2018/5/23
N2 - To obtain effective pedestrian detection results in surveillance video, there have been many methods proposed to handle the problems from severe occlusion, pose variation, clutter background, and so on. Besides detection accuracy, a robust surveillance video system should be stable to video quality degradation by network transmission, environment variation, and so on. In this paper, we conduct the research on the robustness of pedestrian detection algorithms to video quality degradation. The main contribution of this paper includes the following three aspects. First, a large-scale distorted surveillance video data set (DSurVD) is constructed from high-quality video sequences and their corresponding distorted versions. Second, we design a method to evaluate detection stability and a robustness measure called robustness quadrangle, which can be adopted to the visualize detection accuracy of pedestrian detection algorithms on high-quality video sequences and stability with video quality degradation. Third, the robustness of seven existing pedestrian detection algorithms is evaluated by the built DSurVD. Experimental results show that the robustness can be further improved for existing pedestrian detection algorithms. In addition, we provide much in-depth discussion on how different distortion types influence the performance of pedestrian detection algorithms, which is important to design effective pedestrian detection algorithms for surveillance.
AB - To obtain effective pedestrian detection results in surveillance video, there have been many methods proposed to handle the problems from severe occlusion, pose variation, clutter background, and so on. Besides detection accuracy, a robust surveillance video system should be stable to video quality degradation by network transmission, environment variation, and so on. In this paper, we conduct the research on the robustness of pedestrian detection algorithms to video quality degradation. The main contribution of this paper includes the following three aspects. First, a large-scale distorted surveillance video data set (DSurVD) is constructed from high-quality video sequences and their corresponding distorted versions. Second, we design a method to evaluate detection stability and a robustness measure called robustness quadrangle, which can be adopted to the visualize detection accuracy of pedestrian detection algorithms on high-quality video sequences and stability with video quality degradation. Third, the robustness of seven existing pedestrian detection algorithms is evaluated by the built DSurVD. Experimental results show that the robustness can be further improved for existing pedestrian detection algorithms. In addition, we provide much in-depth discussion on how different distortion types influence the performance of pedestrian detection algorithms, which is important to design effective pedestrian detection algorithms for surveillance.
KW - image processing
KW - image quality
KW - Object detection
KW - video signal processing
KW - video surveillance
UR - http://www.scopus.com/inward/record.url?scp=85047649400&partnerID=8YFLogxK
U2 - 10.1109/ACCESS.2018.2840329
DO - 10.1109/ACCESS.2018.2840329
M3 - 文章
AN - SCOPUS:85047649400
SN - 2169-3536
VL - 6
SP - 28890
EP - 28902
JO - IEEE Access
JF - IEEE Access
ER -