SUPPLYING CHORAL WITH ADVANCED TECHNIQUES FOR STUTTERED PEOPLE

Abstract

Stuttering can significantly impact the quality of a person’s life. Stutterers may refrain from speaking and may waste opportunities to make friends, present own ideas, and opinions in public and to be disadvantaged in job interviews. In this paper, we suggest supplying choral or feedback speech with advanced techniques that can contribute to curing people who stutter. The advanced techniques are the Enhanced One-dimensional Local Binary Patterns (EOLBP) and Adapted Multi-Layer Perceptron for Regression (AMLPR). The EOLBP is a useful feature extraction and the AMLPR is a generative classifier. The Fluency Bank (FB) dataset which includes stuttered speech signals is utilized in this study. The best result of a very high accuracy (99.44%) is achieved in this study.