Application of YOLO11 and Long Short-Term Memory Architecture for Exercise Form Evaluation in Weightlifting
DOI:
https://doi.org/10.26905/jtmi.v11i2.16112Keywords:
LSTM, Pose Analysis, Weightlifting, YOLOAbstract
Exercise provides significant benefits for physical health, and weightlifting has become increasingly popular among fitness enthusiasts. However, improper lifting techniques often lead to injuries, discouraging beginners and affecting long-term training consistency. To address this issue, this study proposes a deep learning approach that automatically evaluates weightlifting form through movement classification. The proposed method integrates the YOLO11n-pose algorithm for detecting keypoints from exercise video recordings and the Long Short-Term Memory (LSTM) network for classifying movement types and determining the correctness of form execution. The model achieved a mean average precision of 88.8% using side-view recordings of single- repetition weightlifting exercises. YOLO11n-pose extracts the coordinates of body keypoints, which are converted into joint angle data and analyzed over time using LSTM to identify movement quality based on expert-validated training data. The trained model was implemented into an iOS application called KorForm, developed using FastAPI, to provide real-time feedback for users. The results demonstrate that combining YOLO11n-pose and LSTM effectively supports weightlifting form evaluation and offers a practical solution for promoting safer and more consistent exercise habits.
Downloads
References
[1] World Health Organization (WHO). (2022). Global Status Report on Physical Activity 2022. Geneva, Switzerland: WHO Health Promotion. Retrieved from https://www.who.int/teams/health-promotion/physical-activity/global-status-report-on-physical-activity-2022
[2] Bukhary, H. A., Basha, N. A., Dobel, A. A., Alsufyani, R. M., Alotaibi, R. A., & Almadani, S. H. (2023). Prevalence and pattern of injuries across the weight-training sports. Cureus. https://doi.org/10.7759/cureus.49759
[3] Noteboom, L., Kemler, E., van Beijsterveldt, A. M. C., Hoozemans, M. J. M., van der Helm, F. C. T., & Verhagen, E. A. L. M. (2023). Factors associated with gym-based fitness injuries: A case-control study. JSAMS Plus, 2, 100032.
[4] Fu, H., Gao, J., & Liu, H. (2023). Human pose estimation and action recognition for fitness movements. Computers & Graphics, 116, 418–426. https://doi.org/10.1016/j.cag.2023.09.008
[5] Manivannan, S., Pradhan, Y., Muhammed, Z., Pooja, H., & Bharathi, R. (2024). Automated gym exercise form checker: Deep learning-based pose estimation. In Smart Trends in Computing and Communications: Proceedings of SmartCom 2024, Volume 1 (pp. 71–84). Springer. https://doi.org/10.1007/978-981-97-1320-2
[6] Chen, J., Wang, J., Yuan, Q., & Yang, Z. (2023). CNN-LSTM model for recognizing video-recorded actions performed in a traditional Chinese exercise. IEEE Journal of Translational Engineering in Health and Medicine, 11, 351–359. https://doi.org/10.1109/JTEHM.2023.3282245
[7] Huang, C., Gochoo, M., & Tan, T. (2023). Two-stream architecture using RGB-based ConvNet and pose-based LSTM for video action recognition. In Proceedings of the 2023 15th International Conference on Innovations in Information Technology (IIT) (pp. 127–131). https://doi.org/10.1109/IIT59782.2023.10366415
[8] Kong, Y., et al. (2024). Unlocking the power of LSTM for long-term time series forecasting. Proceedings of the AAAI Conference on Artificial Intelligence, 39(11), 11968–11976. https://doi.org/10.1609/aaai.v39i11.33303
[9] Purnama, B., Erfianto, B., & Wirawan, I. R. (2024). Time series classification of badminton pose using LSTM with landmark tracking. Jurnal Elektronika, Elektromedika, dan Informatika Medis, 7(1), 27–37.
[10] Muriyah, N. M., Sim, J. H., & Yulianto, A. (2024). Evaluating YOLOv5 and YOLOv8: Advancements in human detection. Journal of Information Systems and Informatics, 6(4), 2999–3015. https://doi.org/10.51519/JOURNALISI.V6I4.944
[11] Effendi, Y., Kristian, Y., L. Z. P. C. S. W., & Yutanto, H. (2023). Pemanfaatan Mediapipe body pose estimation dan dynamic time warping untuk pembelajaran Tari Remo. Jurnal Teknologi dan Manajemen Informatika, 9(2), 183–190. https://doi.org/10.26905/JTMI.V9I2.10408
[12] Nainggolan, S. D. A., & Yuadi, I. (2025). Classification of wrist accessories: Advanced watches with logistic regression, SVM, and deep features from Inception V3 and VGG-19. Jurnal Teknologi dan Manajemen Informatika, 11(1), 25–34. https://doi.org/10.26905/JTMI.V11I1.15571
[13] Cinthiya, C., & Oetama, R. S. (2023). Enhancement of coronary heart disease prediction using stacked long short-term memory. Jurnal Teknologi dan Manajemen Informatika, 9(1), 28–36. https://doi.org/10.26905/JTMI.V9I1.9707
[14] Rijal, K. A., Vitianingsih, A. V., Kristyawan, Y., Maukar, A. L., & Wati, S. F. A. (2024). Forecasting model of Indonesia’s oil & gas and non-oil & gas export value using VAR and LSTM methods. Jurnal Teknologi dan Manajemen Informatika, 10(1), 59–69. https://doi.org/10.26905/JTMI.V10I1.13127
[15] Hendrawan, N. D., & Kolandaisamy, R. (2023). A comparative study of YOLOv8 and YOLO-NAS performance in human detection image. Jurnal Teknologi dan Manajemen Informatika, 9(2), 191–201. https://doi.org/10.26905/JTMI.V9I2.12192
Downloads
Published
Issue
Section
License
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:
(1)Â Copyright of the published articles will be transferred to the journal as the publisher of the manuscripts. Therefore, the author confirms that the copyright has been managed by the journal.
(2) Publisher of JTMI: Jurnal Teknologi dan Manajemen Informatika is University of Merdeka Malang.
(3) The copyright follows Creative Commons Attribution–ShareAlike License (CC BY SA): This license allows to Share — copy and redistribute the material in any medium or format, Adapt — remix, transform, and build upon the material, for any purpose, even commercially.
