An Evaluation Approach of Multi-person Movement Synchronization Level using OpenPose

In dance, broadcast gymnastics and other activities, participants are required to keep their action in synchronization, which is very important to show their uniformity. In this paper, we present an approach to quantitatively evaluate the synchronization level of multi-person movement. Firstly, Kinect is used to obtain color image and depth image, and use OpenPose to process color image to acquire 2D joint point coordinates of all participants in the image; Secondly, color image and depth image are fused into point cloud array, and retrieve 3D joint point coordinates corresponding to 2D from point cloud array; Then, according to 3D joint point coordinates, define three joint angles that can represent the human pose; Finally, after smoothing, calculate the evaluation index to achieve the quantitative evaluation. The experimental results demonstrate that when there are fewer than 5 participants, this method can be applied to evaluate the synchronization level of their movements, it has good accuracy and reliability.