Design and research on automatic recognition system of sports dance movement based on computer vision and parallel computing

Abstract

Abstract Professional team sports analyst periodic analysis to obtain strategic and tactical insights into the players and the team s behavior. Target motion analysis team, including systematic determination of hostile team weaknesses, and assess trained team performance and improvement potential. Team video analysis of the current investigation is usually based on workflow. Analysts can also use information visualization techniques to draw the trajectory of the ball and the players. Video analysis is usually a time-consuming process requiring analysts to remember and comment on the scene. In contrast, typically rely on visual data generally used abstract visual mapping model abstraction. They are no longer directly linked to the observed motion context but obtained from the raw tracking data. The proposal provides low-level functions and advanced motion recognition rate by another parallel motion analysis module used as input to evaluate the dancers performance. Application of computer vision techniques to extract the appropriate data from the track of the video input. Besides, advance besides and motion analysis technology can be used to derive and regional analysis, event analysis and correlation analysis of team players motion analysis measurements for football analysis. Our system is a visual mode that allows video and analysts to take advantage of both research forms. Some experts on the team sports analyst survey showed the effectiveness of this integrated approach.