TARU PUBLICATIONS
Journal of Statistics and Management Systems cover

Journal of Statistics and Management Systems

(Since 1998)
Open Access ·Peer-reviewed·ISSN (Online): 2169-0014·ISSN (Print): 0972-0510
Powered by:Powered by

The Journal of Statistics and Management Systems (JSMS) is a world leading journal publishing high quality, rigorously peer-reviewed original research on theoretical and applied statistics and management systems since 1998. The scope is intentionally broad, but papers must make a novel contribution to the field to be considered for publication. Topics include, but are not limited to, the following: • Statistics • Applied Statistics • Industrial Statistics • Statistical Inference • Interdisciplinary role of Statistics • Actuarial Sciences • Decision Sciences • Managerial Aspects • Management Sciences • Management Information Systems

Issues up to 2022 co-published with and available at:Taylor & Francis Online
submissions@tarupublications.com
Taru Online First Current IssueSpecial IssuesBrowse all Issues
Journal of Statistics and Management Systems
Open Access Research Article

Posture recognition in exercise frames via HOBV-based deep sparse learning

* , ,

* Corresponding author · click or hover a name for details

pp. 201–212Vol. 29Issue 2February 2026DOI: 10.47974/JSMS-1565XML
Received:
10 Jun 2025
Published Online:
22 Jan 2026
Article type:
Research Article
Language:
EN
Article no.:
JSMS-1565
Pages:
201–212

Abstract

In the age of machine learning, Human Exercise Recognition (HER) is an area that has been extensively researched. When it comes to computer vision, action recognition is the process of classifying a human exercise that is seen in a video as belonging to one of a selection of prepared actions.  This paper proposes a new method for recognizing physical activity using a Hybrid Object Boundary Value Deep Sparse Auxiliary Network (HOBV-DSAN). The technique combines object boundary value characteristics with deep sparse learning frameworks to improve the depiction of complicated human motions. Our model captures the global structure as well as the fine-grained motion features of physical activities by integrating boundary-aware representations with a sparsity-driven deep auxiliary network. Experimental results show that the proposed technique outperforms traditional models such as YOLO, Alpha Pose, and Deep Pose on many performance criteria such as Accuracy (95.33%), Sensitivity (99.05%), Specificity (97.47%), and Precision (94.57%). The findings demonstrate the HOBV-DSAN model’s durability and accuracy in detecting a broad range of workouts, making it a suitable option for intelligent physical activity monitoring systems.

Keywords

Exercise recognitionSparse coefficientGabor filterDeep learning

Subject Classifications

68T07

References

[1] S. S. Intille and L. Bao, “Physical activity recognition from acceleration data under semi-naturalistic conditions,” Technical Report (2003).
[2] H. Zhang, L. Li, W. Jia, J. D. Fernstrom, R. J. Sclabassi, and M. Sun, “Recognizing physical activity from ego-motion of a camera,” in Proc. IEEE Int. Conf. Eng. Med. Biol. Soc. (EMBS), pp. 5569–5572 (2010), doi: 10.1109/IEMBS.2010.5626794.
[3] E. M. Tapia, S. S. Intille, W. Haskell, K. Larson1, J. Wright, A. King, R. Friedman, “Real-time recognition of physical activities and their intensities using wireless accelerometers and a heart monitor,” in Proc. Int. Symp. Wearable Computers, pp. 37–40 (2007).
[4] L. Li, H. Zhang, W. Jia, J. Nie, W. Zhang, and M. Sun, “Automatic video analysis and motion estimation for physical activity classification,” in Proc. IEEE 36th Annu. Northeast Bioeng. Conf., pp. 1–2 (2010).
[5] Y. Meng, R. Panda, C. C. Lin, P. Sattigeri, L. Karlinsky, K. Saenko, and R. Feris, “AdaFuse: Adaptive temporal fusion network for efficient action recognition,” arXiv preprint, arXiv:2102.05775 (2021).
[6] X. Li, M. Zhang, J. Gu, and Z. Zhang, “Fitness action counting based on MediaPipe,” in Proc. 15th Int. Congr. Image Signal Process., Biomed. Eng. Informatics (CISP-BMEI), Beijing, China, pp. 1–7 (2022), doi: 10.1109/CISPBMEI56279.2022.9980337.
[7] D. Kumar and A. Sinha, Yoga Pose Detection and Classification Using Deep Learning. Saarbrücken, Germany: LAP Lambert Academic Publishing (2020).
[8] G. Taware, R. Kharat, P. Dhende, P. Jondhalekar, and R. Agrawal, “AI-based workout assistant and fitness guide,” in Proc. 6th Int. Conf. Comput., Commun., Control Autom. (ICCUBEA), Pune, India, pp. 1–4 (2022), doi: 10.1109/ICCUBEA54992.2022.10010733.
[9] R. R. Kanase, A. N. Kumavat, R. D. Sinalkar, and S. Somani, “Pose estimation and correcting exercise posture,” in ITM Web Conf., vol. 40, pp. 03031, EDP Sciences (2021).
[10] A. Anilkumar, K. T. Ardra, S. Athulya, K. A. Sarath, and S. Sreeja, “Pose estimated yoga monitoring system,” SSRN Electron. J. (2021), doi: 10.2139/ssrn.3882498.
[11] K. Y. Chen, J. Shin, M. A. M. Hasan, J. J. Liaw, Y. Ouchi, and Y. Tomioka, “Fitness movement types and completeness detection using a transfer-learning-based deep neural network,” Sensors, vol. 22, no. 15, p. 5700 (Jul. 2022), doi: 10.3390/s22155700.
[12] A. S. B. Pauzi, F. B. M. Nazri, S. Sai, A. M. Bataineh, M. N. Hisyam, M. H. Zafar, M. . A. Wahab, A. S. A. Mohamed, “Movement estimation using MediaPipe BlazePose,” in Advances in Visual Informatics (IVIC 2021), H. B. Zaman et al., Eds. Cham, Switzerland: Springer, Lecture Notes in Computer Science, vol. 13051, pp. 574–583 (2021), doi: 10.1007/978-3-030-90235-3_49.
[13] H.-S. Fang, S. Xie, Y.-W. Tai, and C. Lu, “RMPE: Regional multi-person pose estimation,” in Proc. IEEE Int. Conf. Comput. Vis. (ICCV), pp. 2334–2343 (2017).
[14] J. Ding, S. Niu, Z. Nie, and W. Zhu, “Research on human posture estimation algorithm based on YOLO-Pose,” Sensors, vol. 24, no. 10, p. 3036 (2024), doi: 10.3390/s24103036.
[15] A. Toshev and C. Szegedy, “DeepPose: Human pose estimation via deep neural networks,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), pp. 1653–1660 (2014), doi: 10.1109/CVPR.2014.214.

Views: 51Downloads: 7Citations: 0