Computerized System to Aid Deaf Children in Speech Learning.

摘要

This paper describes a voice analyzer, whose purpose is deaf children's assistance in the process of speech learning. The processing of the user's speech signal is performed in real time in order to get an instantaneous feedback of the result of speech training. The aim of this analyzer is not to find the distinction between spoken words, main objective of a speech recognizer but to calculate a level of correctness in the toggle of a specific word, Voice signal analysis was developed through a digital signal processor (DSP), applying spectral analysis processes, extraction of voice's formants, adaptation of formants to the standard levels in domain of time and frequency and statistical matching of the acquired speech signal and the standard one, resulted from training. After calculating the correctness coefficient, the system goes off a visual feedback to the user in the form of a graphical animation, in accordance with the matching ratio, that will determine the progression on speech training.