This study focuses on deep learning classifiers of physical therapy exercises using inertial measurement unit data. Additionally, a comprehensive analysis of the effect of sensor density, location, type, state estimation, and sample size on performance is presented.
GitHub: https://github.com/CMU-MBL/IMU_Exercise_Prediction.
Data: Downloads/View.
Link to the paper: https://doi.org/10.1109/JBHI.2024.3368042.
Code: Instructions can be found at the GitHub link above.
Citation: If you use any of the data or code, please cite the following paper:
V. Phan, K. Song, R. S. Silva, K. G. Silbernagel, J. R. Baxter and E. Halilaj, "Seven Things to Know About Exercise Classification With Inertial Sensing Wearables," in IEEE Journal of Biomedical and Health Informatics, vol. 28, no. 6, pp. 3411-3421, June 2024, doi: 10.1109/JBHI.2024.3368042.