Speech reconstruction using machine learning approach for speech impaired persons
R Kruthika, P. Rajeswari · International journal of advance research, ideas and innovations in technology · 2018
The speech disordered persons are able to produce speech which sounds like they are whispering. The main objective of this work is to reconstruct the abnormal to normal sounding speech by using MFCC coefficients to extract the feature and use these to train cascaded Gaussian Mixture Model (GMM) and Objective measures are used to evaluate the performance of the work. The data used for the work are from WTIMIT online corpus and the speech signals recorded from speech impaired subjects. In this work STRAIGHT toolbox is not employed for its complexity and muffled voice. The obtained SNR is reduced