Performance Analysis for Diving Sport Using YoLoV8, OpenPose and Fuzzy Logic

Diving has been an integral part of the Olympic Games since 1904. The research paper presents an innovative method for evaluating professional diving performance, a task traditionally performed by skilled judges. Through various deep learning techniques and frameworks, we propose an approach for the assessment process, which could eventually eliminate the constant need for manual judgment for athletes. Through a diverse dataset spanning various diving heights and styles, including forward and backward dives, and accounting for gender differences, our approach ensures precise performance classification. Leveraging YoloV8 for dive categorization and Open Pose for accurate pose detection and angle computation, our model enables continuous monitoring of intricate body movements. The integration of Fuzzy Logic further enhances the analysis, encompassing crucial elements such as form, execution, entry technique, and degree of difficulty. This advanced model not only facilitates precise performance evaluation but also allows for meaningful comparisons with elite divers and established benchmarks. In essence, this research marks a paradigm shift in the domain of professional diving performance assessment, equipping coaches and divers alike with powerful tools to refine their skills and elevate the standards of this esteemed sport.