PoseTracker: Accuracy Evaluation of AI-Based Mobile Application for Exercise Posture Feedback
DOI:
https://doi.org/10.21512/ijcshai.v3i1.15123Keywords:
Mobile Application, Computer Vision, MediaPipe, Physical Exercise, Posture DetectionAbstract
In recent years, the rising of public health awareness has increased fitness activities participation. However, improper exercise form remains a significant contributor to injuries, particularly in unsupervised environments. To address this, PoseTracker’s accuracy was evaluated as a native Android application that provides real time feedback on exercise posture through MediaPipe based Human Pose Estimation (HPE) model. The system extracts 33 3D body landmarks, normalizes them to account for body scale, and employs cosine similarity to compare user movements against a reference dataset. Evaluations involving participants aged between 17 to 50 years old and 240 repetitions across four exercises demonstrated high detection accuracy: 88.33% for jumping jacks, 85% for squats, 83.33% for push-ups and 82% for sit ups. While performance can be influenced by environmental factors such as inconsistent lighting, camera positioning and incomplete body visibility, these results highlight the potential for lightweight, AI driven tools to support safe and self-guided fitness routines. Overall, the evaluations indicate that PoseTracker achieves reliable detection accuracy in distinguishing correct and incorrect exercise posture across multiple movement types under realistic conditions. Although performance variability exists due to environmental and system constraints, the accuracy levels observed demonstrate the feasibility of MediaPipe based Human Pose Estimation (HPE) for practical posture assessment in mobile fitness applications.
References
[1] Abhinand G, Mohammed Anas, Naveen Kumar B, Radha G, and Varsha Jituri, “AI Fitness Trainer Using Human,” Dec. 2023, doi: 10.17577/IJERTCONV11IS08017.
[2] D. A. Bonilla et al., “Exercise Selection and Common Injuries in Fitness Centers: A Systematic Integrative Review and Practical Recommendations,” Oct. 01, 2022, MDPI. doi: 10.3390/ijerph191912710.
[3] M. Cuthbertson-Moon, P. A. Hume, H. E. Wyatt, I. Carlson, and B. Hastings, “Gym and Fitness Injuries amongst those Aged 16–64 in New Zealand: Analysis of Ten Years of Accident Compensation Corporation Injury Claim Data,” Sports Med Open, vol. 10, no. 1, Dec. 2024, doi: 10.1186/s40798-024-00694-9.
[4] R. Bouguezzi, S. Sammoud, A. Markov, Y. Negra, and H. Chaabene, “Why Flexibility Deserves to Be Further Considered as a Standard Component of Physical Fitness: A Narrative Review of Existing Insights from Static Stretching Study Interventions,” Youth, vol. 3, no. 1, pp. 146–156, Jan. 2023, doi: 10.3390/youth3010010.
[5] P. Sousa Basto and P. Ferreira, “Mobile applications, physical activity, and health promotion,” BMC Health Serv Res, vol. 25, no. 1, Dec. 2025, doi: 10.1186/s12913-025-12489-z.
[6] H. Dahiya and A. K. Saini, “Assessing individuals’ attitude and behavioural intention to use dietary and fitness mobile applications: evidence from India,” Journal of Indian Business Research, vol. 16, no. 3, pp. 329–352, Jul. 2024, doi: 10.1108/JIBR-12-2022-0302.
[7] H. J. Chae, J.-B. Kim, G. Park, D. M. O’Sullivan, J. Seo, and J.-J. Park, “An Artificial Intelligence Exercise Coaching Mobile App: Development and Randomized Controlled Trial to Verify Its Effectiveness in Posture Correction,” Interact J Med Res, vol. 12, p. e37604, Sep. 2023, doi: 10.2196/37604.
[8] Y. C. Wu, S. X. Lin, J. Y. Lin, C. C. Han, C. S. Chang, and J. X. Jiang, “Development of AI Algorithm for Weight Training Using Inertial Measurement Units,” Applied Sciences (Switzerland), vol. 12, no. 3, Feb. 2022, doi: 10.3390/app12031422.
[9] Y. Zhang, C. Zhao, Y. Yao, C. Wang, G. Cai, and G. Wang, “Human posture estimation and action recognition on fitness behavior and fitness,” Alexandria Engineering Journal, vol. 107, pp. 434–442, Nov. 2024, doi: 10.1016/j.aej.2024.07.039.
[10] H. Kotte, M. Kravčík, and N. Duong-Trung, “Real-Time Posture Correction in Gym Exercises: A Computer Vision-Based Approach for Performance Analysis, Error Classification and Feedback,” 2023. [Online]. Available: http://ceur-ws.org
[11] H. Ji, K. S. Githinji, and T. Kenji, “AI Fitness Coach at Home using Image Recognition,” Sep. 29, 2022. doi: 10.21203/rs.3.rs-2047283/v1.
[12] P. Verma and R. Sharma, “‘Enhancing Yoga Practice through Real-time Posture Detection and Correction using Artificial Intelligence: A comprehensive Review,’” 2023, doi: 10.48047/nq.2023.21.6.NQ23111.
[13] A. Patil, D. Rao, K. Utturwar, T. Shelke, and E. Sarda, “Body Posture Detection and Motion Tracking using AI for Medical Exercises and Recommendation System,” ITM Web of Conferences, vol. 44, p. 03043, 2022, doi: 10.1051/itmconf/20224403043.
[14] S. Saleem, J. Nunes, and M. Aruna, “TrainERAI-Live Gym Tracker using Artificial Intelligence.” [Online]. Available: https://ssrn.com/abstract=4383182
[15] C. Zhen Yue, L. Chee Yong, and L. Nai Shyan, “EXERCISE QUALITY ANALYSIS USING AI MODEL AND COMPUTER VISION,” 2022.
[16] L. Noteboom, E. Kemler, A. M. C. van Beijsterveldt, M. J. M. Hoozemans, F. C. T. van der Helm, and E. A. L. M. Verhagen, “Factors associated with gym-based fitness injuries: A case-control study,” JSAMS Plus, vol. 2, p. 100032, 2023, doi: 10.1016/j.jsampl.2023.100032.
[17] E. I. Wallace, “Repetitive Strain Injury and Electronic Medical records: A brief literature review cjni.net/journal.” [Online]. Available: https://cjni.net/journal/?p=10870
[18] B. A. Bushman, “The Value of Warm-Up and Cool-Down,” 2024. [Online]. Available: www.acsm-healthfitness.org
[19] Y. Sudo, K. Kubo, T. Shinohara, and K. Nakagawa, “Effectiveness of a Warm-up Program with Dynamic Stretching in Preventing Sports Injuries,” Exercise Medicine, vol. 6, p. 1, Mar. 2022, doi: 10.26644/em.2022.001.
[20] D. Sople and R. B. Wilcox, “Dynamic Warm-ups Play Pivotal Role in Athletic Performance and Injury Prevention,” Arthrosc Sports Med Rehabil, 2024, doi: 10.1016/j.asmr.2024.101023.
[21] A. Garbett, Z. Degutyte, J. Hodge, and A. Astell, “Towards Understanding People’s Experiences of AI Computer Vision Fitness Instructor Apps,” in DIS 2021 - Proceedings of the 2021 ACM Designing Interactive Systems Conference: Nowhere and Everywhere, Association for Computing Machinery, Inc, Jun. 2021, pp. 1619–1637. doi: 10.1145/3461778.3462094.
[22] J. L. Chung, L. Y. Ong, and M. C. Leow, “Comparative Analysis of Skeleton-Based Human Pose Estimation,” Future Internet, vol. 14, no. 12, Dec. 2022, doi: 10.3390/fi14120380.
[23] C. Lugaresi et al., “MediaPipe: A Framework for Building Perception Pipelines,” Jun. 2019, [Online]. Available: http://arxiv.org/abs/1906.08172
[24] W. Simoes, L. Reis, C. Araujo, and J. Maia, “Accuracy Assessment of 2D Pose Estimation with MediaPipe for Physiotherapy Exercises,” in Procedia Computer Science, Elsevier B.V., 2024, pp. 446–453. doi: 10.1016/j.procs.2024.11.132.
[25] J. W. Kim, J. Y. Choi, E. J. Ha, and J. H. Choi, “Human Pose Estimation Using MediaPipe Pose and Optimization Method Based on a Humanoid Model,” Applied Sciences (Switzerland), vol. 13, no. 4, Feb. 2023, doi: 10.3390/app13042700.
[26] L. Gupta, S. Gurbuxani, and K. Madan, “Virtual Fitness Trainer using Artificial Intelligence,” in Proceedings of the 2024 Sixteenth International Conference on Contemporary Computing, New York, NY, USA: ACM, Aug. 2024, pp. 226–233. doi: 10.1145/3675888.3676056.
[27] Y. Kwon and D. Kim, “Real-Time Workout Posture Correction using OpenCV and MediaPipe,” The Journal of Korean Institute of Information Technology, vol. 20, no. 1, pp. 199–208, Jan. 2022, doi: 10.14801/jkiit.2022.20.1.199.
[28] S. A. Lohi, S. S. Katwate, O. M. Ladole, I. D. Kusumbe, and K. S. Jaminkar, “Realtime Pose Estimation using AI,” 2025. [Online]. Available: www.isteonline.in
[29] R. Riccio, “Real-Time Fitness Exercise Classification and Counting from Video Frames,” 2024.
[30] E. Bagga and A. Yang, “Real-Time Posture Monitoring and Risk Assessment for Manual Lifting Tasks Using MediaPipe and LSTM,” in Proceedings of the 1st International Workshop on Multimedia Computing for Health and Medicine, New York, NY, USA: ACM, Oct. 2024, pp. 79–85. doi: 10.1145/3688868.3689199.
[31] K. Y. Chen, J. Shin, M. A. M. Hasan, J. J. Liaw, O. Yuichi, and Y. Tomioka, “Fitness Movement Types and Completeness Detection Using a Transfer-Learning-Based Deep Neural Network,” Sensors, vol. 22, no. 15, Aug. 2022, doi: 10.3390/s22155700.
[32] M. Latyshev, G. Lopatenko, V. Shandryhos, O. Yarmoliuk, M. Pryimak, and I. Kvasnytsia, “COMPUTER VISION TECHNOLOGIES FOR HUMAN POSE ESTIMATION IN EXERCISE: ACCURACY AND PRACTICALITY,” SOCIETY. INTEGRATION. EDUCATION. Proceedings of the International Scientific Conference, vol. 2, pp. 626–636, May 2024, doi: 10.17770/sie2024vol2.7842.
[33] S. Dill et al., “Accuracy Evaluation of 3D Pose Estimation with MediaPipe Pose for Physical Exercises,” in Current Directions in Biomedical Engineering, Walter de Gruyter GmbH, Sep. 2023, pp. 563–566. doi: 10.1515/cdbme-2023-1141.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Billy Collhins, Kalyana Mitta, Christian Gunawan, Sonya Rapinta Manalu
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License - Share Alike that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
USER RIGHTS
All articles published Open Access will be immediately and permanently free for everyone to read and download. We are continuously working with our author communities to select the best choice of license options, currently being defined for this journal as follows: Creative Commons Attribution-Share Alike (CC BY-SA)
