Exploiting Hierarchical Dense Structures on Hypergraphs for Multi-Object Tracking

Longyin Wen, Zhen Lei, Siwei Lyu, Stan Z. Li, Ming Hsuan Yang

Research output: Contribution to journalArticle

25 Citations (Scopus)

Abstract

Most multi-object tracking algorithms are developed within the tracking-by-detection framework that consider the pairwise appearance similarities between detection responses or tracklets within a limited temporal window, and thus less effective in handling long-term occlusions or distinguishing spatially close targets with similar appearance in crowded scenes. In this work, we propose an algorithm that formulates the multi-object tracking task as one to exploit hierarchical dense structures on an undirected hypergraph constructed based on tracklet affinity. The dense structures indicate a group of vertices that are inter-connected with a set of hyperedges with high affinity values. The appearance and motion similarities among multiple tracklets across the spatio-temporal domain are considered globally by exploiting high-order similarities rather than pairwise ones, thereby facilitating distinguish spatially close targets with similar appearance. In addition, the hierarchical design of the optimization process helps the proposed tracking algorithm handle long-term occlusions robustly. Extensive experiments on various challenging datasets of both multi-pedestrian and multi-face tracking tasks, demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods.

Original languageEnglish
Article number7360186
Pages (from-to)1983-1996
Number of pages14
JournalIEEE transactions on pattern analysis and machine intelligence
Volume38
Issue number10
DOIs
Publication statusPublished - 2016 Oct 1

Fingerprint

Object Tracking
Hierarchical Structure
Hypergraph
Occlusion
Affine transformation
Pairwise
Face Tracking
Target
Process Optimization
Higher Order
Motion
Demonstrate
Experiment
Similarity
Experiments

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Vision and Pattern Recognition
  • Computational Theory and Mathematics
  • Artificial Intelligence
  • Applied Mathematics

Cite this

@article{8a24bdbf41e8470283ffe1d93b0ae17c,
title = "Exploiting Hierarchical Dense Structures on Hypergraphs for Multi-Object Tracking",
abstract = "Most multi-object tracking algorithms are developed within the tracking-by-detection framework that consider the pairwise appearance similarities between detection responses or tracklets within a limited temporal window, and thus less effective in handling long-term occlusions or distinguishing spatially close targets with similar appearance in crowded scenes. In this work, we propose an algorithm that formulates the multi-object tracking task as one to exploit hierarchical dense structures on an undirected hypergraph constructed based on tracklet affinity. The dense structures indicate a group of vertices that are inter-connected with a set of hyperedges with high affinity values. The appearance and motion similarities among multiple tracklets across the spatio-temporal domain are considered globally by exploiting high-order similarities rather than pairwise ones, thereby facilitating distinguish spatially close targets with similar appearance. In addition, the hierarchical design of the optimization process helps the proposed tracking algorithm handle long-term occlusions robustly. Extensive experiments on various challenging datasets of both multi-pedestrian and multi-face tracking tasks, demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods.",
author = "Longyin Wen and Zhen Lei and Siwei Lyu and Li, {Stan Z.} and Yang, {Ming Hsuan}",
year = "2016",
month = "10",
day = "1",
doi = "10.1109/TPAMI.2015.2509979",
language = "English",
volume = "38",
pages = "1983--1996",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "10",

}

Exploiting Hierarchical Dense Structures on Hypergraphs for Multi-Object Tracking. / Wen, Longyin; Lei, Zhen; Lyu, Siwei; Li, Stan Z.; Yang, Ming Hsuan.

In: IEEE transactions on pattern analysis and machine intelligence, Vol. 38, No. 10, 7360186, 01.10.2016, p. 1983-1996.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Exploiting Hierarchical Dense Structures on Hypergraphs for Multi-Object Tracking

AU - Wen, Longyin

AU - Lei, Zhen

AU - Lyu, Siwei

AU - Li, Stan Z.

AU - Yang, Ming Hsuan

PY - 2016/10/1

Y1 - 2016/10/1

N2 - Most multi-object tracking algorithms are developed within the tracking-by-detection framework that consider the pairwise appearance similarities between detection responses or tracklets within a limited temporal window, and thus less effective in handling long-term occlusions or distinguishing spatially close targets with similar appearance in crowded scenes. In this work, we propose an algorithm that formulates the multi-object tracking task as one to exploit hierarchical dense structures on an undirected hypergraph constructed based on tracklet affinity. The dense structures indicate a group of vertices that are inter-connected with a set of hyperedges with high affinity values. The appearance and motion similarities among multiple tracklets across the spatio-temporal domain are considered globally by exploiting high-order similarities rather than pairwise ones, thereby facilitating distinguish spatially close targets with similar appearance. In addition, the hierarchical design of the optimization process helps the proposed tracking algorithm handle long-term occlusions robustly. Extensive experiments on various challenging datasets of both multi-pedestrian and multi-face tracking tasks, demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods.

AB - Most multi-object tracking algorithms are developed within the tracking-by-detection framework that consider the pairwise appearance similarities between detection responses or tracklets within a limited temporal window, and thus less effective in handling long-term occlusions or distinguishing spatially close targets with similar appearance in crowded scenes. In this work, we propose an algorithm that formulates the multi-object tracking task as one to exploit hierarchical dense structures on an undirected hypergraph constructed based on tracklet affinity. The dense structures indicate a group of vertices that are inter-connected with a set of hyperedges with high affinity values. The appearance and motion similarities among multiple tracklets across the spatio-temporal domain are considered globally by exploiting high-order similarities rather than pairwise ones, thereby facilitating distinguish spatially close targets with similar appearance. In addition, the hierarchical design of the optimization process helps the proposed tracking algorithm handle long-term occlusions robustly. Extensive experiments on various challenging datasets of both multi-pedestrian and multi-face tracking tasks, demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods.

UR - http://www.scopus.com/inward/record.url?scp=84986328532&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84986328532&partnerID=8YFLogxK

U2 - 10.1109/TPAMI.2015.2509979

DO - 10.1109/TPAMI.2015.2509979

M3 - Article

AN - SCOPUS:84986328532

VL - 38

SP - 1983

EP - 1996

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 10

M1 - 7360186

ER -