Optical-Flow and Human-Segmentation Based Method for High Accuracy Detection of Golf Swing Key Events

Golf swing sequencing is a crucial aspect of golf performance analysis, requiring precise identification of key events throughout the swing motion. Traditional methods for analyzing golf swings face limitations in accuracy and efficiency, prompting the exploration of AI-driven solutions. In this paper, we propose a novel method that integrates golfers’ movement obtained from human segmentation with optical flow, along with whole frame analysis to enhance the accuracy of key event detection in golf swing videos. The proposed approach utilizes a pipeline consisting of human segmentation, optical flow, feature extraction, and classification to detect key events in the golf swing. Additionally, various data augmentation techniques are employed to enrich the training dataset. Therefore, our research improves model generalization and performance. Experimental results demonstrate the effectiveness of our proposed method, achieving a Percentage of Correct Events (PCE) of 82.2% on the GolfDB dataset, a promising result compared to SwingNet.