Online Classification Method of Sports Videos Based on Wavelet Transform and SVM Algorithm

Abstract

This study introduces an innovative online classification methodology for sports video content employing a combination of wavelet transform and the Support Vector Machine (SVM) algorithm. In an era where sports video content is proliferating at an unprecedented rate, the imperative for efficient online classification has become ever more critical to augment the viewing experience and fulfill commercial objectives. This methodology harnesses the robust time-frequency feature extraction capabilities of wavelet transform alongside the formidable classification prowess of the SVM algorithm, culminating in an automated system adept at discerning various sports and competition actions. The efficacy of our approach is substantiated through extensive large-scale experimentation, with results evidencing commendable accuracy and real-time performance metrics in the realm of online sports video classification. Furthermore, this research delves into the strengths and potential constraints of our method, proffering avenues for prospective inquiry. The contributions of this study are twofold. It offers novel solutions for video processing and enhances the online viewership experience in the sports domain, thus holding substantial import for it.