Decomposition of MES using wavelet transform and support vector machine

This paper involves the development of a semiautomatic system for the decomposition from myoelectric signals (MES) in their essential components, motor unit action potential (MUAP). Using tools of analysis (wavelet analysis) and classification (support vector machines) for signal processing. The purpose of the decomposition is to obtain the largest number of MUAPs and their features to evaluate the performance of the muscle. It has healthy intramuscular records, sampled at 10 kHz for 10 seconds. The MUAP are selected by calculating the wavelet transform. With the help of the sampling rate and the duration of potentials, some MUAPs are selected to serve as patterns of each family in the classification. Using support vector machines the MUAPs are classified in the different families resulting from the previous step. Finally, firing rates are calculated for each family.