Chained-Tracker: Chaining Paired Attentive Regression Results for End-to-End Joint Multiple-Object Detection and Tracking

被引:282
作者
Peng, Jinlong [1 ]
Wang, Changan [1 ]
Wan, Fangbin [2 ]
Wu, Yang [3 ]
Wang, Yabiao [1 ]
Tai, Ying [1 ]
Wang, Chengjie [1 ]
Li, Jilin [1 ]
Huang, Feiyue [1 ]
Fu, Yanwei [2 ]
机构
[1] Tencent Youtu Lab, Shanghai, Peoples R China
[2] Fudan Univ, Shanghai, Peoples R China
[3] Nara Inst Sci & Technol, Ikoma, Japan
来源
COMPUTER VISION - ECCV 2020, PT IV | 2020年 / 12349卷
关键词
Multiple-object Tracking; Chained-Tracker; End-to-end solution; Joint detection and tracking;
D O I
10.1007/978-3-030-58548-8_9
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Existing Multiple-Object Tracking (MOT) methods either follow the tracking-by-detection paradigm to conduct object detection, feature extraction and data association separately, or have two of the three subtasks integrated to form a partially end-to-end solution. Going beyond these sub-optimal frameworks, we propose a simple online model named Chained-Tracker (CTracker), which naturally integrates all the three subtasks into an end-to-end solution (the first as far as we know). It chains paired bounding boxes regression results estimated from overlapping nodes, of which each node covers two adjacent frames. The paired regression is made attentive by object-attention (brought by a detection module) and identity-attention (ensured by an ID verification module). The two major novelties: chained structure and paired attentive regression, make CTracker simple, fast and effective, setting new MOTA records on MOT16 and MOT17 challenge datasets (67.6 and 66.6, respectively), without relying on any extra training data. The source code of CTracker can be found at: github.com/pjl1995/CTracker.
引用
收藏
页码:145 / 161
页数:17
相关论文
共 39 条
[1]   Confidence-Based Data Association and Discriminative Deep Appearance Learning for Robust Online Multi-Object Tracking [J].
Bae, Seung-Hwan ;
Yoon, Kuk-Jin .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2018, 40 (03) :595-610
[2]   Tracking without bells and whistles [J].
Bergmann, Philipp ;
Meinhardt, Tim ;
Leal-Taixe, Laura .
2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, :941-951
[3]   Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics [J].
Bernardin, Keni ;
Stiefelhagen, Rainer .
EURASIP JOURNAL ON IMAGE AND VIDEO PROCESSING, 2008, 2008 (1)
[4]  
Bochinski E, 2017, 2017 14TH IEEE INTERNATIONAL CONFERENCE ON ADVANCED VIDEO AND SIGNAL BASED SURVEILLANCE (AVSS)
[5]   Soft-NMS - Improving Object Detection With One Line of Code [J].
Bodla, Navaneeth ;
Singh, Bharat ;
Chellappa, Rama ;
Davis, Larry S. .
2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2017, :5562-5570
[6]   Robust Tracking-by-Detection using a Detector Confidence Particle Filter [J].
Breitenstein, Michael D. ;
Reichlin, Fabian ;
Leibe, Bastian ;
Koller-Meier, Esther ;
Van Gool, Luc .
2009 IEEE 12TH INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2009, :1515-1522
[7]   Enhancing Detection Model for Multiple Hypothesis Tracking [J].
Chen, Jiahui ;
Sheng, Hao ;
Zhang, Yang ;
Xiong, Zhang .
2017 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW), 2017, :2143-2152
[8]   Factors Influencing Pediatric Emergency Department Visits for Low-Acuity Conditions [J].
Long, Christina M. ;
Mehrhoff, Casey ;
Abdel-Latief, Eman ;
Rech, Megan ;
Laubham, Matthew .
PEDIATRIC EMERGENCY CARE, 2021, 37 (05) :265-268
[9]   Near-Online Multi-target Tracking with Aggregated Local Flow Descriptor [J].
Choi, Wongun .
2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2015, :3029-3037
[10]   FAMNet: Joint Learning of Feature, Affinity and Multi-dimensional Assignment for Online Multiple Object Tracking [J].
Chu, Peng ;
Ling, Haibin .
2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019), 2019, :6171-6180