Design of Dance Movement Recognition Algorithm Based on 3D Motion Capture Data

Although the traditional photography, audio recording and video recording methods are easy to collect and make, they can't record the dancers' body movements in detail, let alone make scientific analysis and research on them and apply them to choreography. This paper proposes a dance movement Recognition technology based on 3D motion capture data. After fully considering the characteristics of dance movements, the HOG (Histogram of Oriented Gradient) features extracted from dance videos are used to represent dance movements. In this paper, the image with HOG feature extraction has included the edge feature information of the target in all the images in the video segment after accumulation operation. The time series HMM (Hidden Markov model) is used to study the spatial information of adjacent joint points in the human skeleton model. Experiments show that the correct recognition rate of the proposed method can reach 97.797% when the number of nodes is 20. Experiments show that compared with one-dimensional Markov model method, the method proposed in this paper has better recognition effect and stronger anti-interference ability.