Optimization of Table Tennis Players' Technical Movements based on Genetic Algorithm

Abstract

Table tennis players employ rapid footwork, accurate racket movements, and strategic ball placement. They use topspin, backspin, and side spin to control the ball, adjusting their posture and grip accordingly. To improve table tennis players' presentation, we suggested an inventive Redefined Genetic Algorithm (RGA) for optimizing their industrial movements and refining sports play strategy efficiently. By competently identifying edges in imagery, our technique can reduce error and speed of negative response while growing identification accuracy. Furthermore, this study builds a tactical analysis framework for table tennis video games by combining the techniques of trajectory prediction and tracking of targets. We obtained a dataset that comprises table tennis player performance metrics, gameplay videos, and trajectory data. Image processing is performed to pre-process the gathered raw image data. Utilizing Scale-Invariant Feature Transform (SIFT) for efficient feature extraction enhances the robustness and accuracy of identifying key points and descriptors. Long Short-Term Memory (LSTM) and Radial Basis Function Support Vector Machine (RBF-SVM) algorithms were utilized to recognize and analyse movement patterns in table tennis gameplay videos effectively. The proposed model is implemented in Python software. We evaluate the effectiveness of the proposed method using various key metrics such as precision, recall, f1 score, and accuracy. The experimental findings demonstrate the effectiveness of the proposed recognition model.