Intelligent Motion Capture Technology of Dance Pose Analysis Using Target Deduction and Gaussian Mixture Model

In order to achieve intelligent capture of dance posture and better analyse dance movements. The objective is to execute the distinctive stands that the ultra-high Kinect sensor detects for each dance, based on information about the body joints. Based on the enhanced Gaussian mixture model (GMM), the implemented Target Deduction (TD) algorithm, which face based on Gaussian mixture model. 6 traditional dance routines and their variants or included in the used data set. That training data set has been optimised for the intended use of the gesture recognition technique, which makes use of several time limitations, spatial information, and spatial distribution features. For target dancing areas with fewer background photo station under motor damage conditions, the Gaussian Mixed Model distribution background model account, the algorithm, and their frame difference can be extracted in straight lines. Both of these techniques successfully identify dancing moving picture targets with their Gaussian Mixture Model (GMM)-based realtime Target Detection (TD) algorithm. The implemented method achieves 98.95% of accuracy, which compared to state-of-the-art methods.