Real-time attitude estimation motion analysis method and system, computer equipment and storage medium

Abstract

The invention discloses a real-time attitude estimation motion analysis method and system, computer equipment and a storage medium. The method comprises the following steps: acquiring a real-time video of a user; inputting the video frame into a trained double-branch deep network for feature acquisition to obtain a joint point heat map of a human body and an affinity region between joint points; performing non-maximum suppression on multiple peak values of the joint point thermograph, selecting and obtaining a series of candidate joint points, connecting the candidate joint points with one another to form a bipartite graph, and optimizing the bipartite graph; according to the optimized bipartite graph, performing distortion correction on joint pixel points between adjacent video frames when the user moves in real time, calculating limb angle information, and obtaining limb movement data; and after a consulting instruction sent by the user is received, performing motion analysis on thelimb motion data, and outputting a motion analysis result. According to the invention, the image space features can be more effectively reserved, and unnecessary influence when optimal connection is found between the established joint points is eliminated.

Description

technical field [0001] The invention relates to a real-time attitude estimation motion analysis method, system, computer equipment and storage medium, belonging to the field of computer vision. Background technique [0002] With the implementation of the national fitness program and more and more people participating in various sports, the analysis of human body movement is particularly important. In recent years, with the rapid development of big data and artificial intelligence technology, more and more people use mobile phone videos, applications, etc. as guidance, and wear various expensive equipment to help them correct their exercise posture. Analysis cannot better express the quality of motion, so the method of pose estimation is proposed to analyze the motion standard level of motion video, which provides more space and possibility for the development of motion analysis. [0003] Pose estimation belongs to the category of computer vision, and specifically refers to ...

AI Technical Summary

Problems solved by technology

The shortcomings of the current algorithm and the current situation: (1) High-precision sensors are required to collect motion information, resulting in increased motion costs; (2) the algorithm can only complete the pose estimation of a single person, and cannot analyze multiple people and multiple types of motion at the same time ; (3) Action analysis that cannot satisfy real-time performance

[0004] Traditional sports analysis methods such as coaches, wearable sports equipment, and teaching video guidance, on the one hand, may increase the cost of sports, on the other hand, they can only conduct one-to-one analysis, cannot analyze multiple people at the same time, and the accuracy is mostly low

With the introduction of pose estimation algorithms such as DeepPose, MoDeep, DeepCut, and DeeperCut, the research on single-person pose estimation technology in images has become increasingly mature. Due to the overlapping parts of multiple people in the video, the above algorithms are almost unable to correctly estimate the joints of different people. There are still few studies on how to connect and combine human body pose estimation into the analysis of video human motion data. Therefore, how to realize multi-person pose estimation and combine estimated data for motion analysis is a key problem that needs to be solved.

Images