A Method of Yoga Action Pattern Recognition Based on Computer Vision Technology

Object detection and recognition using computer vision has been a very interesting and challenging research field in the past three decades. Classification and target location based on machine learning and computer vision technology has always been a hot topic, and great achievements have been made. Firstly, the classification analysis method is used to analyze and study the yoga movement and theory based on computer vision technology, including human eye tracking and recognition, face recognition, head movement tracking and recognition, gesture recognition and posture recognition. Based on the task of yoga movement pattern recognition, a yoga movement pattern recognition method based on computer vision technology is proposed. The improved model is based on the framework and structure of the network. A certain number of candidate regions are proposed and classified through feature extraction, and then these regions are output as the detected bounding box. The posture action diagram of yoga, that is, ordinary RGB color image, is collected by RGB camera, and the bone data is extracted from RGB image by bone extraction model. The RGB image data and bone data are input into the joint model, and the joint model will output the category of Yoga action and the score of this action.