An efficient human detection method for multi-pedestrian tracking

Abstract

Traditional human detection using pre-trained detectors tends to be computationally intensive for time-critical tracking tasks, and the detection rate is prone to be unsatisfying when occlusion, motion blur and body deformation occur frequently. A spatial-confidential proposal filtering method (SCPF) is proposed for efficient and accurate human detection. It consists of two filtering phases: spatial proposal filtering and confidential proposal filtering. A compact spatial proposal is generated in the first phase to minimize the search space to reduce the computation cost. The human detector only estimates the confidence scores of the candidate search regions accepted by the spatial proposal instead of global scanning. At the second phase, each candidate search region is assigned with a supplementary confidence score according to their reliability estimated by the confidential proposal to reduce missing detections. The performance of the SCPF method is verified by extensive tests on several video sequences from available public datasets. Both quantitatively and qualitatively experimental results indicate that the proposed method can highly improve the efficiency and the accuracy of human detection.

Keywords

human detection / spatial proposal filtering / confidential proposal filtering

References

[1] DalalN, TriggsB. Histograms of oriented gradients for human detection [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2005PiscatawayIEEE886-893

[2] WuJ, GeyerC, RehgJ. Real-time human detection using contour cues [C]. Proceedings of the IEEE International Conference on Robotics and Automation, 2011PiscatawayIEEE860-867

[3] FelzenszwalbP F, GirshickR B, McallesterD, RamananD. Object detection with discriminatively trained part based models [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(9): 1627-1645

[4] ChenY-q, RuiY, HuangT. JPDAF based HMM for real-time contour tracking [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2001PiscatawayIEEEI-543

[5] CoxI, HingoraniS. An efficient implementation of Reid’s multiple hypothesis tracking algorithm and its evaluation for the purpose of visual tracking [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1996, 18(2): 138-150

[6] OhS, RussellS, SastryS. Markov chain Monte Carlo data association for multi-target tracking [J]. IEEE Transactions on Automatic Control, 2009, 54(3): 481-497

[7] SmithK, Gatica-perezD, OdobezJ. Using particles to track varying numbers of interacting people [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2005PiscatawayIEEE962-969

[8] HaritaogluI, HarwoodD, DavisL. W4: Real-time surveillance of people and their activities [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(8): 809-830

[9] NilliusP, SullivanJ, CarlssonS. Multi-target tracking linking identities using bayesian network inference [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2006PiscatawayIEEE2187-2194

[10] SongX, CuiJ-s, ZhaH-b, ZhaoH-jing. Vision-based multiple interacting targets tracking via on-line supervised learning [C]. Computer Vision-ECCV 2008, 2008BerlinSpringer642-655

[11] LuH-c, JiaC-h, ZhangR-juan. An effective method for detection and segmentation of the body of human in the view of a single stationary camera [C]. Proceedings of the International Conference on Pattern Recognition, 2009PiscatawayIEEE1-4

[12] GaoT, LiuZ-g, ZhangJun. Traffic video based multiple moving targets tracking [J]. Journal of Central South University: Science and Technology, 2010, 40(3): 1028-1036

[13] BoseB, WangX-g, GrimsonE. Multi-class object tracking algorithm that handles fragmentation and grouping [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2007PiscatawayIEEE1-8

[14] ZhaoT, NevatiaR, WuB. Segmentation and tracking of multiple humans in crowded environments [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2007, 30(7): 1198-1211

[15] KoJ H, KimE S. Stereoscopic video surveillance system for detection of targets 3-D location coordinates and moving trajectories [J]. Optics Communications, 2006, 266(1): 67-79

[16] ViolaP, JonesM. Rapid object detection using a boosted cascade of simple features [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2001PiscatawayIEEEI-511

[17] LiY, ZhouA-h, YamashitaT, KawadeM. Tracking in low frame rate video: A cascade particle filter with discriminative observers of different life spans [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008, 30(10): 1728-1740

[18] StalderS, GrabnerH, VanG L. Cascaded confidence filtering for improved tracking-by-detection [C]. Computer Vision-ECCV 2010, 2010BerlinSpringer369-382

[19] JiaL, TongX-f, LiW-l, WangT, ZhangY-m, WangH-qi. Automatic player detection, labeling and tracking in broadcast soccer video [J]. Pattern Recognition Letters, 2009, 30(2): 103-113

[20] BreitensteinM, ReichlinF, LeibeB, Koller-meierE, VanG L. Robust tracking-by-detection using a detector confidence particle filter [C]. Proceedings of the IEEE International Conference on Computer Vision, 2009PiscatawayIEEE1515-1522

[21] MitzelD, HorbertE, EssA, LeibeB. Multi-person tracking with sparse detection and continuous segmentation [C]. Computer Vision-ECCV 2010, 2010BerlinSpringer397-410

[22] LuW, TingJ, MurphyK, LittleJ J. Identifying players in broadcast sports videos using conditional random fields [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2011PiscatawayIEEE3249-3256

[23] RyooM, AggarwalJ. Observe-and-explain: A new approach for multiple hypotheses tracking of humans and objects [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2008PiscatawayIEEE1-8

[24] BreitensteinM, ReichlinF, LeibeB, Koller-MeierE, Van-GoolL. Online multi-person tracking-by-detection from a single, uncalibrated camera [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(9): 1820-1833

[25] WangF-s, ZhaoQ-jie. A new particle filter for nonlinear filtering problems [J]. Chinese Journal of Computers, 2008, 31(2): 346-352

[26] OkumaK, TaleghaniA, FreitasN, LittleJ J, LoweD G. A boosted particle filter: Multitarget detection and tracking [C]. Computer Vision-ECCV 2004, 2004BerlinSpringer28-39

[27] GordonN, SalmondD, SmithA. Novel approach to nonlinear/non-Gaussian Bayesian state estimation [C]. IEE Proceedings of Radar and Signal Processing, 1993HertfordshireIET107-113

[28] ChengY-zong. Mean shift, mode seeking, and clustering [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1995, 17(8): 790-799

[29] HuangC, WuB, NevatiaR. Robust object tracking by hierarchical association of detection responses [C]. Proceedings of the 10th European Conference on Computer Vision, 2008BerlinSpringer788-801

[30] GrabnerH, BischofH. On-line boosting and vision [C]. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2006PiscatawayIEEE260-267

[31] Lids dataset for AVSS 2007 [EB/OL]. http://www.eecs.qmul.ac.uk/~andrea/avss2007_d.html.

[32] EC Funded CAVIAR project/IST 2001 37540 [EB/OL]. http://homepages.inf.ed.ac.uk/rbf/CAVIAR/.

[33] PETS 2009 S2.L1 PEOPLE TRACKING [EB/OL]. http://www.cvg.rdg.ac.uk/PETS2009/a.html

[34] CannyJ. A computational approach to edge detection [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1986, 6: 679-698

[35] ChangC C, LinC J. LIBSVM: A library for support vector machines [J]. ACM Transactions on Intelligent Systems and Technology, 2011, 2(27): 1-27