Keywords
Pose Estimation
movenet
Machine Learning
XGBoost
How to Cite
Abstract
Klasifikasi fase gerakan pada latihan deadlift penting untuk menghindari cedera akibat postur yang salah. Penelitian ini mengembangkan sistem klasifikasi posisi “up” dan “down” menggunakan estimasi pose berbasis MoveNet Lightning dan algoritma XGBoost. Sistem mendeteksi 17 keypoint tubuh secara real-time melalui browser dan menghitung tiga sudut utama (pinggul, lutut, punggung) sebagai fitur klasifikasi. Data sudut tersebut dilabeli otomatis dan dilatih menggunakan XGBoost dengan akurasi uji mencapai 98,51%. Analisis menunjukkan bahwa sudut lutut memiliki kontribusi tertinggi terhadap hasil klasifikasi. Sistem berjalan sepenuhnya di sisi klien menggunakan React.js dan TensorFlow.js, memungkinkan klasifikasi real-time tanpa backend. Pendekatan ini menunjukkan bahwa kombinasi model ringan dan algoritma efisien dapat menghasilkan sistem pelatih digital yang akurat, ringan, dan portabel.
References
[1] TensorFlow Hub, “MoveNet: Ultra fast and accurate pose detection model,” https://www.tensorflow.org/hub/tutorials/movenet.
[2] C. Arrowsmith, D. Burns, T. Mak, M. Hardisty, and C. Whyne, “Physiotherapy Exercise Classification with Single-Camera Pose Detection and Machine Learning,” Sensors, vol. 23, no. 1, p. 363, Dec. 2022, doi: 10.3390/s23010363.
[3] A. ÇELİKEL and I. A. OZKAN, “Deep Learning-Based Classification of Powerlifting Movements Using Mediapipe,” International Journal of Applied Methods in Electronics and Computers, Sep. 2024, doi: 10.58190/ijamec.2024.107.
[4] C. Arrowsmith, D. Burns, T. Mak, M. Hardisty, and C. Whyne, “Physiotherapy Exercise Classification with Single-Camera Pose Detection and Machine Learning,” Sensors, vol. 23, no. 1, p. 363, Dec. 2022, doi: 10.3390/s23010363.
[5] C. Mercadal-Baudart et al., “Exercise quantification from single camera view markerless 3D pose estimation,” Heliyon, vol. 10, no. 6, p. e27596, Mar. 2024, doi: 10.1016/j.heliyon.2024.e27596.
[6] B. Liang, W. Qin, and Z. Liao, “A Differential Evolutionary-Based XGBoost for Solving Classification of Physical Fitness Test Data of College Students,” Mathematics, vol. 13, no. 9, p. 1405, Apr. 2025, doi: 10.3390/math13091405.
[7] T. Chen and C. Guestrin, “XGBoost,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York, NY, USA: ACM, Aug. 2016, pp. 785–794. doi: 10.1145/2939672.2939785.
[8] W. Wahid, A. A. AlArfaj, E. A. Alabdulqader, T. Sadiq, H. Rahman, and A. Jalal, “Advanced Human Pose Estimation and Event Classification Using Context-Aware Features and XGBoost Classifier,” IEEE Access, vol. 12, pp. 179839–179856, 2024, doi: 10.1109/ACCESS.2024.3498093.
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Copyright (c) 2025 Kevin Ragil Krisna Dyansyah, Ahmad Bagus Setiawan, Patmi Kasih