Action Recognition and Temporal Localization for Wushu Routine Competition
Kaiqi Zhang, Yiqun Pang, Qiurui Wang, Fengmei Li · 2022
To pursue the unification and standardization of technical actions, promote the standardization of the wushu competition scoring system, and help wushu enter the Olympic Games, this paper adopts a method based on computer vision algorithms for action recognition and localization of the actions of the international long fist routine v3, and the results show that the application of this method can better slice and dice action clips from the original video. Combining this method with the competition scoring rules, it is promising to create an automatic scoring referee system to automate the scoring of martial arts routine competitions.