EMG-Based Hand Gestures Classification Using Machine Learning Algorithms

Abstract

Most chronic disorders in the human motor system occur due to functional defects in the nervous system. In this regard, prostheses have been invented to help rehabilitate people with sensory-motor disorders, providing a suitable environment for improving and restoring motor function. These assistive devices attempt to help affected people reach normal mobility and attain a sense of touch for those paralyzed or missing limbs. Myoelectric control is used in prosthetic systems to allow users to control the movement of the prosthesis using the muscle electrical signals produced by muscle contraction, called electromyography (EMG). Classification methods in controlling assistive devices can significantly contribute to identifying specific movements or muscle functions based on the EMG signals. In recent years, Machine Learning techniques have become increasingly popular for classifying EMG signals due to their ability to recognize patterns in the data and classify them with high accuracy. In this work, we present a classification model based on the Artificial Neural Network (ANN) to classify four hand gestures: neutral, point, hook, and lateral pinch. We compared the proposed ANN model with other Machine Learning algorithms, such as LongShort-Term-Memory (LSTM), K-Nearest Neighbor (KNN), Naïve Bayes, CatBoost, Support Vector Machine (SVM), and Random Forest. The results demonstrated that the ANN model was superior to other models, with an accuracy of 93% in classifying EMG signals.