A speech recognition system based on electromyography for the rehabilitation of dysarthric patients: A Thai syllable study

The objective of this study is to develop a speech recognition system for classifying nine Thai syllables, which is used for the rehabilitation of dysarthric patients, based on five channels of surface electromyography (sEMG) signals from the human articulatory muscles. After the sEMG signal from each channel was collected, it was processed by a band-pass filter from 20–450Hz for noise removal. Then, six features from three feature categories were determined and analyzed, namely, mean absolute value (MAV) and wavelength (WL) from amplitude based features (ABF), zero crossing (ZC) and mean frequency (MNF) from frequency based features (FBF), and L-kurtosis (L-KURT) and L-skewness (L-SKW) from statistics based features (SBF). Subsequently, a spectral regression extreme learning machine (SRELM) was used as the feature projection technique to reduce the dimension of feature vector from 30 to 8. Finally, the projected features were classified using a feed forward neural network (NN) classifier with 5-fold cross-validation. The proposed system was evaluated with the sEMG signals from seven healthy volunteers and five dysarthric volunteers. The results show that the proposed system can recognize the sEMG signals from both healthy and dysarthric volunteers. The average classification accuracies obtained from all six features in the healthy and dysarthric volunteers were 94.5% and 89.4%, respectively.