2023 Vol.6 Aug.N04

• Title: An One-Shot Multi-Object tracking Methods Based on Multi-Layer Feature Fusion Mechanism
• Name: Boyi Wang, Xiaopeng Hu, Xinrong Wu, Fan Wang
• Company: Dalian University of Technology ,Dalian,116000 China
• Abstract:

  Due to balanced accuracy and speed, one-shot models which jointly learn detection and identifification embeddings, have drawn great attention in multi-object tracking (MOT). However, the inherent differences and relationships between detection and re identification (ReID) were unintentionally overlooked, as in the one-shot tracking paradigm, they were seen as two isolated tasks. This results in inferior performance compared with existing two-stage methods. To overcome decreased feature accuracy in single-stage multi-object tracking algorithms, this paper proposes a feature fusion-based MOT method to improve feature extraction accuracy.  Experimental results demonstrate that the proposed method effectively solves the problem of feature accuracy reduction in multi-object tracking tasks and exhibits good performance.

  
  

• Keyword: Multi-Ojbect Tracking; One-Shot; Attention; RE-ID
• DOI: 10.12250/jpciams2023090614
• Citation form: Boyi Wang.An One-Shot Multi-Object tracking Methods Based on Multi-Layer Feature Fusion Mechanism [J]. Computer Informatization and Mechanical System,2023,Vol.6,pp.64-65

[1] ZHANG Y, SUN P, JIANG Y, et al. Bytetrack: Multi-object tracking by associating every detection box[C]. Computer Vision–ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XXII. Cham: Springer Nature Switzerland, 2022: 1-21.

[2] WANG Z, ZHENG L, LIU Y, et al. Towards real-time multi-object tracking[C]. Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XI 16. Springer International Publishing, 2020: 107-122.

[3] ZHOU X, KOLTUN V, KRÄHENBÜHL P. Tracking objects as points[C]. Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part IV. Cham: Springer International Publishing, 2020: 474-490.

[4] ZHANG Y, WANG C, WANG X, et al. Fairmot: On the fairness of detection and re-identification in multiple object tracking[J]. International Journal of Computer Vision, 2021, 129: 3069-3087.

[5] J WU, J CAO, L SONG, et al, Track to Detect and Segment: An Online Multi-Object Tracker, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021: 12347-12356.

[6] SUN P, CAO J, JIANG Y, et al. Transtrack: Multiple object tracking with transformer[J]. arXiv preprint arXiv:2012.15460, 2020.

[7] PANG J, QIU L, LI X, et al. Quasi-dense similarity learning for multiple object tracking[C]. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition(CVPR), 2021: 164-173.

[8] SUN, PEI. DanceTrack: Multi-Object Tracking in Uniform Appearance and Diverse Motion[C]. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021: 20961-20970.