Review of Recent Development in EMG Signal Classification Methods

Characterizing electromyographic (EMG) signals is essential in diagnosing neuromuscular diseases. Different classifiers have been developed to build accurate and highly efficient EMG signal characterization algorithms. This paper reviews some classification approaches used in EMG characterization and discusses the most recent achievements in this field, emphasizing neuromuscular disease. The methods studied are classified according to their methodology, and a synopsis of the significant findings related to each way is provided. Surface EMG is recently more relative to signal processing and artificial intelligence for moving organs. On the other hand, traditional EMG is under the scope of surgical cases by following different algorithms. The specified algorithm is our interest to start work from the endpoint of a light spot like Wavelet Transform (WT), Discrete Fourier transform (DFT), Higher order statistics (HOS), Artificial Neural Network (ANN), Fuzzy Logic (FL), Empirical Mode Decomposition (EMD), Independent Component Analysis (ICA), Principle Component Analysis (PCA) and Autoregressive (AR).