Fencing Action Recognition System Using Deep Learning and KNN Algorithm

This article proposes a design scheme for a fencing action recognition system based on deep learning. The system uses mediapipe Pose model and OpenCV model to extract human skeleton angle information, and uses KNN algorithm for human posture recognition. In terms of data cleaning, this article uses data preprocessing and data balancing techniques. The system interface uses the tkinter standard GUI library, the programming language is Python, and the IDE is PyCharm. Experimental results show that the system can accurately recognize fencing actions. This scheme can provide posture recognition and technical analysis support for fencing athletes.