sEMG Signals Classification using CNN Features Extraction as a Reliable Method
sEMG (Surface electromyography) signals are essential in several applications, such as in prosthetic control. These signals are collected and analyzed to produce the expected actions through corresponding pattern recognition. In this sense, feature extraction plays a critical role in achieving good accuracy during the classification process. In recent years, with the interesting results obtained through convolutional filters and supervised learning, it is possible to extract properties that best distinguish and classify images. Therefore, this research work uses a CNN network to extract these features that will be later applied in conventional classifiers. The obtained results allowed to verify that the proposed methodology guarantees better results when compared to the works that use traditional characteristics for the classification process.