On Pairwise Costs for Network Flow Multi-Object Tracking

被引:0
|
作者
Chari, Visesh [1 ,2 ]
Lacoste-Julien, Simon [3 ]
Laptev, Ivan [2 ]
Sivic, Josef [2 ]
机构
[1] INRIA, Paris, France
[2] Ecole Normale Super, WILLOW Project Team, Dept Informat, ENS INRIA CNRS UMR 8548, Paris, France
[3] Ecole Normale Super, SIERRA Project Team, Dept Informat, ENS INRIA CNRS UMR 8548, Paris, France
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Multi-object tracking has been recently approached with the min-cost network flow optimization techniques. Such methods simultaneously resolve multiple object tracks in a video and enable modeling of dependencies among tracks. Min-cost network flow methods also fit well within the "tracking-by-detection" paradigm where object trajectories are obtained by connecting per-frame outputs of an object detector. Object detectors, however, often fail due to occlusions and clutter in the video. To cope with such situations, we propose to add pairwise costs to the min-cost network flow framework. While integer solutions to such a problem become NP-hard, we design a convex relaxation solution with an efficient rounding heuristic which empirically gives certificates of small suboptimality. We evaluate two particular types of pairwise costs and demonstrate improvements over recent tracking methods in real-world video sequences.
引用
收藏
页码:5537 / 5545
页数:9
相关论文
共 50 条
  • [1] Deep Network Flow for Multi-Object Tracking
    Schulter, Samuel
    Vernaza, Paul
    Choi, Wongun
    Chandraker, Manmohan
    [J]. 30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, : 2730 - 2739
  • [2] Learning of Global Objective for Network Flow in Multi-Object Tracking
    Li, Shuai
    Kong, Yu
    Rezatofighi, Hamid
    [J]. 2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2022, : 8845 - 8855
  • [3] Structure and Appearance Preserving Network Flow for Multi-object Tracking
    Pu, Shi
    Zhang, Honggang
    Zhao, Kaili
    [J]. 2016 23RD INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2016, : 1804 - 1808
  • [4] Multi-object tracking based on network flow model and ORB feature
    Jieyu Chen
    Zhenghao Xi
    Junxin Lu
    Jingjing Ji
    [J]. Applied Intelligence, 2022, 52 : 12282 - 12300
  • [5] Multi-object tracking based on network flow model and ORB feature
    Chen, Jieyu
    Xi, Zhenghao
    Lu, Junxin
    Ji, Jingjing
    [J]. APPLIED INTELLIGENCE, 2022, 52 (11) : 12282 - 12300
  • [6] Data Association with Graph Network for Multi-Object Tracking
    Wu, Yubin
    Sheng, Hao
    Wang, Shuai
    Liu, Yang
    Ke, Wei
    Xiong, Zhang
    [J]. KNOWLEDGE SCIENCE, ENGINEERING AND MANAGEMENT, PT I, 2022, 13368 : 268 - 280
  • [7] Associative affinity network learning for multi-object tracking
    Ma, Liang
    Zhong, Qiaoyong
    Zhang, Yingying
    Xie, Di
    Pu, Shiliang
    [J]. FRONTIERS OF INFORMATION TECHNOLOGY & ELECTRONIC ENGINEERING, 2021, 22 (09) : 1194 - 1206
  • [8] MULTI-OBJECT TRACKING BY VIRTUAL NODES ADDED MIN-COST NETWORK FLOW
    Liu, Peixin
    Li, Xiaofeng
    Feng, Haoyang
    Fu, Zhizhong
    [J]. 2017 24TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2017, : 2577 - 2581
  • [9] Multi-object trajectory tracking
    Han, Mei
    Xu, Wei
    Tao, Hai
    Gong, Yihong
    [J]. MACHINE VISION AND APPLICATIONS, 2007, 18 (3-4) : 221 - 232
  • [10] Referring Multi-Object Tracking
    Wu, Dongming
    Han, Wencheng
    Wang, Tiancai
    Dong, Xingping
    Zhang, Xiangyu
    Shen, Jianbing
    [J]. 2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2023, : 14633 - 14642