MSMCT: Multi-State Multi-Camera Tracker

Behzad BOZORGTABAR, Roland GOECKE

    Research output: Contribution to journalArticle

    Abstract

    Visual tracking of multiple persons simultaneously is an important tool for group behaviour analysis. In this paper, we demonstrate that multi-target tracking in a network of nonoverlapping cameras can be formulated in a framework, where the association among all given target hypotheses both within and between cameras is performed simultaneously. Our approach helps to overcome the fragility of multi-camera based tracking, where the performance relies on the single-camera tracking results obtained at input level. In particular, we formulate an estimation of the target states as a multi-state graph optimisation problem, in which the likelihood of each target hypothesis belonging to different identities is modelled. In addition, we learn the target-specific model to improve the similarity measure among targets based on the appearance cues. We also handle the occluded targets when there is no reliable evidence for the target’s presence and each target trajectory is expected to be fragmented into multiple tracks. An iterative procedure is proposed to solve the optimisation problem, resulting in final trajectories that reveal the true states of the targets. The performance of the proposed approach has been extensively evaluated on challenging multi-camera non-overlapping tracking datasets, in which many difficulties such as occlusion, viewpoint and illumination variation are present. The results of systematic experiments conducted on a large set of sequences show that the proposed approach outperforms several state-of-the-art trackers.
    Original languageEnglish
    Article number8048028
    Pages (from-to)3361-3376
    Number of pages16
    JournalIEEE Transactions on Circuits and Systems for Video Technology
    Volume28
    Issue number12
    Early online date21 Sep 2017
    DOIs
    Publication statusE-pub ahead of print - 21 Sep 2017

    Fingerprint

    Cameras
    Trajectories
    Target tracking
    Lighting
    Experiments

    Cite this

    @article{e1c3eccee91f406a89687a902024bb15,
    title = "MSMCT: Multi-State Multi-Camera Tracker",
    abstract = "Visual tracking of multiple persons simultaneously is an important tool for group behaviour analysis. In this paper, we demonstrate that multi-target tracking in a network of nonoverlapping cameras can be formulated in a framework, where the association among all given target hypotheses both within and between cameras is performed simultaneously. Our approach helps to overcome the fragility of multi-camera based tracking, where the performance relies on the single-camera tracking results obtained at input level. In particular, we formulate an estimation of the target states as a multi-state graph optimisation problem, in which the likelihood of each target hypothesis belonging to different identities is modelled. In addition, we learn the target-specific model to improve the similarity measure among targets based on the appearance cues. We also handle the occluded targets when there is no reliable evidence for the target’s presence and each target trajectory is expected to be fragmented into multiple tracks. An iterative procedure is proposed to solve the optimisation problem, resulting in final trajectories that reveal the true states of the targets. The performance of the proposed approach has been extensively evaluated on challenging multi-camera non-overlapping tracking datasets, in which many difficulties such as occlusion, viewpoint and illumination variation are present. The results of systematic experiments conducted on a large set of sequences show that the proposed approach outperforms several state-of-the-art trackers.",
    keywords = "multi-camera target tracking, multiple states graphical model, target-specific metric learning",
    author = "Behzad BOZORGTABAR and Roland GOECKE",
    year = "2017",
    month = "9",
    day = "21",
    doi = "10.1109/TCSVT.2017.2755038",
    language = "English",
    volume = "28",
    pages = "3361--3376",
    journal = "IEEE Transactions on Circuits and Systems for Video Technology",
    issn = "1051-8215",
    publisher = "IEEE, Institute of Electrical and Electronics Engineers",
    number = "12",

    }

    MSMCT: Multi-State Multi-Camera Tracker. / BOZORGTABAR, Behzad; GOECKE, Roland.

    In: IEEE Transactions on Circuits and Systems for Video Technology, Vol. 28, No. 12, 8048028, 12.2028, p. 3361-3376.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - MSMCT: Multi-State Multi-Camera Tracker

    AU - BOZORGTABAR, Behzad

    AU - GOECKE, Roland

    PY - 2017/9/21

    Y1 - 2017/9/21

    N2 - Visual tracking of multiple persons simultaneously is an important tool for group behaviour analysis. In this paper, we demonstrate that multi-target tracking in a network of nonoverlapping cameras can be formulated in a framework, where the association among all given target hypotheses both within and between cameras is performed simultaneously. Our approach helps to overcome the fragility of multi-camera based tracking, where the performance relies on the single-camera tracking results obtained at input level. In particular, we formulate an estimation of the target states as a multi-state graph optimisation problem, in which the likelihood of each target hypothesis belonging to different identities is modelled. In addition, we learn the target-specific model to improve the similarity measure among targets based on the appearance cues. We also handle the occluded targets when there is no reliable evidence for the target’s presence and each target trajectory is expected to be fragmented into multiple tracks. An iterative procedure is proposed to solve the optimisation problem, resulting in final trajectories that reveal the true states of the targets. The performance of the proposed approach has been extensively evaluated on challenging multi-camera non-overlapping tracking datasets, in which many difficulties such as occlusion, viewpoint and illumination variation are present. The results of systematic experiments conducted on a large set of sequences show that the proposed approach outperforms several state-of-the-art trackers.

    AB - Visual tracking of multiple persons simultaneously is an important tool for group behaviour analysis. In this paper, we demonstrate that multi-target tracking in a network of nonoverlapping cameras can be formulated in a framework, where the association among all given target hypotheses both within and between cameras is performed simultaneously. Our approach helps to overcome the fragility of multi-camera based tracking, where the performance relies on the single-camera tracking results obtained at input level. In particular, we formulate an estimation of the target states as a multi-state graph optimisation problem, in which the likelihood of each target hypothesis belonging to different identities is modelled. In addition, we learn the target-specific model to improve the similarity measure among targets based on the appearance cues. We also handle the occluded targets when there is no reliable evidence for the target’s presence and each target trajectory is expected to be fragmented into multiple tracks. An iterative procedure is proposed to solve the optimisation problem, resulting in final trajectories that reveal the true states of the targets. The performance of the proposed approach has been extensively evaluated on challenging multi-camera non-overlapping tracking datasets, in which many difficulties such as occlusion, viewpoint and illumination variation are present. The results of systematic experiments conducted on a large set of sequences show that the proposed approach outperforms several state-of-the-art trackers.

    KW - multi-camera target tracking

    KW - multiple states graphical model

    KW - target-specific metric learning

    UR - https://ieeexplore.ieee.org/document/8048028

    UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-85030630505&origin=resultslist&sort=plf-f&src=s&st1=MSMCT%3a+Multi-State+Multi-Camera+Tracker&st2=&sid=8f0b2bfdaa207bea86e2768e1b21488c&sot=b&sdt=b&sl=46&s=TITLE%28MSMCT%3a+Multi-State+Multi-Camera+Tracker%29&relpos=0&citeCnt=0&searchTerm=

    U2 - 10.1109/TCSVT.2017.2755038

    DO - 10.1109/TCSVT.2017.2755038

    M3 - Article

    VL - 28

    SP - 3361

    EP - 3376

    JO - IEEE Transactions on Circuits and Systems for Video Technology

    JF - IEEE Transactions on Circuits and Systems for Video Technology

    SN - 1051-8215

    IS - 12

    M1 - 8048028

    ER -