Taijiquan training method and system based on dynamic recognition

Abstract

The invention provides a dynamic recognition-based shadowboxing training method and system, and relates to the technical field of video processing. The shadowboxing training method based on dynamic recognition comprises the following steps: acquiring real-time human body motion image data through image acquisition equipment, and creating a forwarding data packet for the acquired human body motion image data; based on a skeleton node recognition algorithm, recognizing a skeleton node image corresponding to the target user contained in the human body motion image data, and analyzing and calculating skeleton key point data of a motion state; and acquiring standard skeleton key point data as a key skeleton node image model, outputting key skeleton node images in sequence, and comparing the key skeleton node images with real-time human body motion image data acquired by the image acquisition equipment to judge whether the target user completes a standard action sequence or not. When the user performs action learning, the difficulty is reduced, the interestingness is increased, the enthusiasm is enhanced, and the practice effect is improved.

Description

technical field [0001] The present application relates to the technical field of video processing, and in particular, to a Tai Chi training method and system based on dynamic recognition. Background technique [0002] Human motion pose recognition is a process of detecting the positions of human skeleton joints given an image or a video, and classifying and labeling human poses according to the structural features of the joints. The detection of human skeleton joints is a key step in human pose recognition. With the development of deep learning technology, the detection effect of human skeleton joints has been continuously improved, and it has been applied in the related fields of computer vision, which has attracted the attention of researchers. [0003] The patent "Pedestrian Detection Method Based on Video Surveillance" (CN201010227766.5) uses the extended gradient histogram feature and the Adaboost algorithm to quickly detect pedestrians, and then uses the gradient hist...

AI Technical Summary

Problems solved by technology

[0004] There are also some video demonstrations of coherent movements in Rope Skipping every day, such as group dance troupe, radio exercises, etc. However, the demonstration videos and body images belong to different areas, and users need to compare the actions they have made with the demonstration videos during the practice process. Consistent, for continuous movements, it is almost impossible for the user to make an accurate instant comparison during the movement

[0005] These research results mainly focus on the algorithms of pedestrian detection and event detection, using traditional machine learning methods for feature extraction and detection and recognition, but they have not been involved in the field of national sports, such as the single picture of the Douyin Gadance machine, The user cannot learn coherent movements well, or cannot judge whether his movements are standard during the motor learning process

Images