Tracking algorithm when managing competitive activities of top level teams online based on computer visioncomputer vision
https://doi.org/10.21822/2073-6185-2024-51-2-120-127
Abstract
Objective. The article presents the results of a study of tracking algorithms for analyzing a basketball game. The purpose of the work is to determine the optimal method for using athlete tracking when used online.
Method. The research is based on methods and algorithms for solving management problems in organizational systems.
Result. Algorithms with object re-identification are considered, taking into account both motion dynamics and appearance. The most popular tracking algorithms, BYTE, taken from the Bytetrack algorithm, and the Deepsort algorithm, which showed high results in the task of tracking pedestrians in a crowd, were selected as candidates. The algorithms were compared using the MOTA and MOTP tracking assessment quality metrics, as well as the operating time of the algorithms. The experiments were carried out on a general and sports dataset - MOT20 и SportMot.
Conclusion. The study showed that the best result in online frame processing is achieved by the ByteTrack algorithm. It showed comparable quality metrics with fast turnaround times. The authors used open implementations of the algorithms and wrote a convenient interface for conducting experiments on different datasets and detection sources.
About the Authors
A. A. Polozov
Ural Federal University
Russian Federation
Russian Federation
Andrey A. Polozov, Dr. Sci. (Pedagogical)
19 Mira St., Ekaterinburg 620014
N. A. Maltceva
Ural Federal University
Russian Federation
Russian Federation
Natalya A. Maltceva, Engineer
19 Mira St., Ekaterinburg 620014
G. S. Kramarenko
Ural Federal University
Russian Federation
Russian Federation
Georgy S. Kramarenko, Student
19 Mira St., Ekaterinburg 620014
M. A. Lipilin
Ural Federal University
Russian Federation
Russian Federation
Matvey A. Lipilin, Student
19 Mira St., Ekaterinburg 620014
A. R. Akhmetzyanov
Surgut State Pedagogical University
Russian Federation
Russian Federation
Artur R. Akhmetzyanov, Teacher
10/2 50 years of the Komsomol St., Surgut 628417
References
1. Yifu Zhang1, Peize Sun2, Yi Jiang3, Dongdong Yu3, Fucheng Weng1, Zehuan Yuan3, Ping Luo2, Wenyu Liu1, and Xinggang Wang. ByteTrack: Multi-Object Tracking by Associating Every Detection Box/ Huazhong University of Science and Technology 2 The University of Hong Kong 3 ByteDance Inc.
2. N. Wojke, A. Bewley and D. Paulus, “Simple online and realtime tracking with a deep association metric,” 2017 IEEE International Conference on Image Processing (ICIP), Beijing, China, 2017; 3645-3649, doi: 10.1109/ICIP.2017.8296962.
3. Milan, Anton & Leal-Taixé, Laura & Reid, Ian & Roth, Stefan. (2016). MOT16: A Benchmark for MultiObject Tracking.
4. Bernardin, K., Stiefelhagen, R. Evaluating Multiple Object Tracking Performance: The CLEAR MOT Metrics. J Image Video Proc 2008, 246309 (2008). https://doi.org/10.1155/2008/246309
5. Punn NS, Sonbhadra SK, Agarwal S, Rai G. Monitoring COVID-19 social distancing with person detection and tracking via fine-tuned YOLO v3 and Deepsort techniques. arXiv; 2021.
6. Wang J. et al. SportsTrack: An Innovative Method for Tracking Athletes in Sports Scenes //arXiv preprint arXiv:2211.07173. 2022.
7. M.I.H. Azhar, F.H.K. Zaman, N.M. Tahir and H. Hashim, "People Tracking System Using DeepSORT," 2020 10th IEEE International Conference on Control System, Computing and Engineering (ICCSCE), Penang, Malaysia, 2020, pp. 137-141, doi: 10.1109/ICCSCE50387.2020.9204956.
8. Y. Du et al., “StrongSORT: Making DeepSORT Great Again,” in IEEE Transactions on Multimedia, doi: 10.1109/TMM.2023.3240881.
9. Yutao Cui, Chenkai Zeng, Xiaoyu Zhao, Yichun Yang, Gangshan Wu, and Limin Wang. Sportsmot: A large multi-object tracking dataset in multiple sports scenes. arXiv preprint arXiv:2304.05170, 2023.
10. Patrick Dendorfer, Hamid Rezatofighi, Anton Milan, Javen Shi, Daniel Cremers, Ian D. Reid, Stefan Roth, Konrad Schindler, Laura Leal-Taixé:MOT20: A benchmark for multi object tracking in crowded scenes. CoRR abs/2003.09003 (2020)
11. Ivanov, Yu. S. Analysis of neural network algorithms for object tracking / Yu. S. Ivanov // Manufacturing technologies of the future: from creation to implementation: Proceedings of the V International Scientific and Practical Conference, Komsomolsk-on-Amur, December 06–11, 2021 / Editorial Board: S.I. Sukhorukov (ed.), A.S. Gudim, N.N. Lyubushkina. – Komsomolsk-on-Amur: Komsomolsk-on-Amur State University, 2022;34-37. – EDN LXOWWD. ( In Russ)
12. D.A. Vrazhnov, A.V. Shapovalov, V.V. Nikolaev, “On the quality of operation of algorithms for tracking objects on video,” Computer Research and Modeling, 4:2 (2012), 303–313( In Russ)
13. Chen, H. (2023). Appearance Awarned Detector for MOT: An Enhanced ReID Branch for Tracking Memorize. Academic Journal of Science and Technology, 5(1), 46–48. https://doi.org/10.54097/ajst.v5i1.5345 Electronic resources https://habr.com/ru/articles/514450/
14. https://habr.com/ru/articles/166693/
Review
For citations:
Polozov A.A., Maltceva N.A., Kramarenko G.S., Lipilin M.A., Akhmetzyanov A.R. Tracking algorithm when managing competitive activities of top level teams online based on computer visioncomputer vision. Herald of Dagestan State Technical University. Technical Sciences. 2024;51(2):120-127. (In Russ.) https://doi.org/10.21822/2073-6185-2024-51-2-120-127
Views:
170
Email this article 
