A Kinect-based golf swing classification system using HMM and Neuro-Fuzzy

摘要

This study displays a method of scoring time-sequential postures of golf swing. Correct posture of golf swing is the most important skill for golfer training. In this paper, firstly, a game controller, Kinect, is used to capture the 3D skeleton coordination of a golfer while swing is performed. Secondly, the time-sequential posture of golf swing features has been extracted. Thirdly, a HMM-NF model is used for scoring, which combines ability of HMM model for temporal data modeling with that of Fuzzy Neural Network for fuzz rule modeling and fuzzy defined in a fuzzy (I am not sure on this!!!). Results have shown that the proposed methods can be implemented to identify and score the golf swing effectively with up to 80% accuracy rate.