We designed a human motion statistics system, which is used to count users’ physical activity. The system is based on the recognition of human action, which depends on 3D human pose estimation. We combine the probability model of 3D human pose with the multi-layer CNN architecture to obtain the 3D human pose estimation. Then, different actions are defined by the proportion of distance between each joint, and the amount of action is counted. The total calories burned during exercise are obtained by calculating the calories consumed in each action. Our system is simple and convenient, and can be used in many situations. The experimental results show that our system achieved the desired results.
CITATION STYLE
Huo, Y., Zou, L., Gao, M., Wang, J., Guo, X., & Sun, J. (2019). A System for Calculating the Amount of Motion Based on 3D Pose Estimation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11633 LNCS, pp. 466–476). Springer Verlag. https://doi.org/10.1007/978-3-030-24265-7_40
Mendeley helps you to discover research relevant for your work.