An Automatic Scoring System for Wushu Movements based on Real-time Action Image Recognition and Matching Algorithm
Li Deng, Hengru Liu · 2023
Existing engineering practices show that it is a complex task to require automatic detection and matching of corresponding features, usually without any prior knowledge of the scene and image. As the detailed application, this research study considered the scheme of automatic scoring system for Wushu movements based on real-time action image recognition and matching algorithm. Starting from the review of the state-of-the-art study results, the questions for the efficient image recognition is firstly presented. Then, efficient effective moment invariants are designed to serve as the pre-processing of the images and the quadtree method is used for the modelling. The new roles and methods are designed to conduct the task of the efficient image recognition. Furthermore, the SIFT based matching algorithm is designed. Combining these 2 algorithms, the efficient scoring system is implemented and comparison experiment is conducted. After testing, it is reflected that the proposed model has the better recognition accuracy and the scoring performance is good.