ALS Disease Detection in EMG using a Polynomial Compact Support Kernel-based Time-Frequency Distribution and CNN

Electromyography (EMG) is an important diagnostic tools used for neuromuscular disorders' classification. Amyotrophic Lateral Sclerosis (ALS) is recognized as one of the most severe neurological disorders. In this paper, our proposed approach involves utilizing time-frequency analysis of EMG signals combined with convolutional neural networks (CNNs) to achieve the most effective classification of ALS and normal EMG signals. Specifically, the Polynomial Cheriet-Belouchrani distribution (PCBD) is herein used to provide the TF signatures related to the analyzed EMG samples. The obtained RGB images are fed to a pre-trained CNN model to perform classification. The obtained results show that the PCBD/CNN-based classifier achieves the best performance compared to other state-of-the-art classifiers, attaining an average test accuracy of 99%.