Dance movement design based on computer three-dimensional auxiliary system

Abstract

The Marching Cube method, a classical algorithm based on the voxel method, is used to perform three-dimensional computer-aided dance movements. Aiming at several problems in the MC algorithm, an optimization and improvement method is proposed, which realizes the organic combination of image segmentation and the MC algorithm. We use the segmented binary data to extract the isosurface; solve the ambiguity problem by establishing a look-up table of intersection conditions; use the method of quickly finding the boundary cube to avoid searching for a certain amount of empty cells; at the same time, use the midpoint method to calculate the triangle vertices. This paper verifies the effectiveness of the dance action recognition algorithm proposed in this paper on the dance data set DanceDB of the University of Cyprus and the folk dance data set FolkDance produced by the laboratory. The experimental results show that the algorithm in this paper can maintain a certain recognition rate for more complex dance movements, and the method in this paper can still guarantee a certain accuracy rate when the background and the target are easily confused. This also verifies the effectiveness of the motion recognition algorithm in this paper for dance motion recognition.