An automatic evaluation method of human body movements and a dance scoring system

Abstract

The invention discloses an automatic evaluation method of human body movements and a dance scoring system. The method includes selecting human skeleton nodes and human contour nodes based on depth data; constructing human limb vectors based on human skeleton nodes, calculating the space angle between the human limb vectors and the corresponding template limb vectors, weighting them to one, and calculating the cumulative error As the difference degree based on the analysis of human skeleton nodes; construct the human body contour vector based on the human body contour node, calculate the space angle between two adjacent human body contour vectors, and use the difference value between the space angle between it and the template contour vector to construct an energy function, The graph cut method is used to obtain the minimum value of the energy function as the degree of difference based on the analysis of human body contour nodes; the weighted sum of the difference degrees based on the analysis of human skeleton nodes and human body contour nodes is used as an evaluation parameter to measure the degree of matching between human body movements and template movements. In order to realize the automatic evaluation of human motion.

Description

technical field [0001] The invention relates to motion analysis technology in the field of image processing and computer vision, in particular to an automatic evaluation method of human body movements and a dance scoring system based on depth data. Background technique [0002] In the application fields of intelligent dance teaching, sports video analysis and rehabilitation training, the automatic evaluation method of human movement has a broad application prospect. Compared with the traditional human motion evaluation method based on human judgment, the automatic evaluation method of human motion will not be affected by human subjective factors, and can intelligently evaluate the captured human motion and the prescribed template motion based on the computer and the preset analysis method the degree of difference between. In the prior art, most automatic evaluation methods of human body motions are based on two-dimensional graphic sequences to calculate the human skeleton a...

AI Technical Summary

Problems solved by technology

This method estimates the human skeleton nodes based on the depth data, and often uses fitting and other methods to estimate the human skeleton nodes twice. Due to the noise in the depth data, there will be some errors in the estimation results, especially for the end nodes of human limbs that describe subtle movements. The estimation results are inaccurate, which makes the accuracy of the evaluation results low

Although the human body motion analysis method based on human body contour information has high accuracy in evaluating subtle movements such as the waist of the human body, it is easily disturbed by human clothing when extracting human body contour nodes in the early stage.

Images