Optimization and Application of Gesture Classification Algorithm Based on EMG

Abstract

Electromyographic signals (EMG) have a wide range of applications in biomedicine, intelligent human-computer interaction, and rehabilitation assistance. However, due to noise interference, individual differences, computational complexity, and other issues, traditional classification strategies still face challenges in terms of accuracy, stability, and real-time performance. In this paper, the traditional EMG gesture classification methods are comprehensively investigated, and data enhancement, feature mining and lightweight deep learning models are used to improve their accuracy and real-time performance. Moreover, multiple modes are combined to optimize boundary processing to improve real-time performance and robustness. Based on various published databases and practical application environments, the proposed optimization algorithm has demonstrated significant improvements across multiple evaluation metrics, showing both high feasibility and practicality. Meanwhile, the possibility of applying the algorithm to intelligent prosthetics, intelligent human-computer interaction, rehabilitation assistance and other fields is discussed, so as to promote the progress of EMG gesture recognition technology.