An Efficient CNN-Based Deep Learning Model for Hand Gesture Recognition Using EMG Signals

This paper presents an efficient CNN-based deep learning model for hand gesture recognition using electromyography (EMG) signals, aimed at improving the accuracy and reliability of human–machine interaction systems. The proposed approach focuses on extracting discriminative spatial–temporal features from preprocessed EMG signals collected from forearm muscles during different hand gestures. A lightweight convolutional neural network architecture is designed to automatically learn complex muscle activation patterns without relying on handcrafted features, thereby reducing computational complexity while maintaining high recognition performance. Noise filtering, normalization, and signal segmentation are applied to enhance signal quality and ensure robustness across users. Experimental evaluation on benchmark EMG gesture datasets demonstrates that the proposed CNN model achieves superior classification accuracy, faster convergence, and better generalization compared to traditional machine learning and existing deep learning methods. The results confirm the suitability of the proposed framework for real-time applications such as prosthetic control, rehabilitation systems, and gesture-based interactive devices.