Leveraging Support Vector Machine for Sports Injury Classification

This study explores the utility of Support Vector Machines (SVM) for the classification of injuries in Athletes including tennis, table tennis, and badminton. The primary objective of this research was to develop an automated method to detect between normal and injury pose, providing a poten-tially powerful tool for early detection. In other words, this research is the pioneer study to represent one another solution for sports injury classification using the image data apart from the current approaches that might be focused on the sensor-based. Therefore, we collect a dataset from various social media across the three sports, labeled into two categories: “Normal” and “Injury”. The SVM model was trained on this dataset, and tuned to optimize its performance in distinguishing the two classes. Achieving an accuracy of 96.63%, the result demonstrates that the SVM model can effectively identify injury patterns in the considered sports. The implications of this study are sig-nificant, presenting the initial steps towards applying computer vision tasks in sports topics. By demonstrating the feasibility of an SVM-based approach to injury detection in these racket sports.