Posture Estimation & Posture Sequence Recognition for Martial Artists

Human body pose estimation is a technology for locating the key points of the human body, and has played an important role in many industries and sports. To date, the method of using this technology to locate the key points of the human body is mainly using a regression heat map. There is no good way to make full use of the structural characteristics of the human body and relationship between key points. Faced with the above difficulties, this article solves the problem by studying the posture estimation and posture sequence of martial artists under the grid convolutional coding neural network. Firstly, use the grid to divide the image to roughly locate the key points of the human body, and then accurately locate the key points of the human body from the offset position output by the grid. This positioning method makes the size of the heat map larger than the grid size so as to achieve high precision positioning and cut down the computation of a convolutional neural network. Under the multi-person scenario in actual martial arts sports, we put forward a multi-person posture estimation algorithm which uses the connection between key points of the human body.

Automated summary

Human body pose estimation is a technology used to locate the key points of the human body in various industries and sports. This paper proposes a method that utilizes grid convolutional coding neural network to address the limitations faced when using regression heat maps for key point localization. By dividing the image into grids and accurately locating the key points, high precision positioning is achieved, and a multi-person posture estimation algorithm is proposed for scenarios involving multiple individuals.