Malay computer aided speech therapy using audio-fingerprint and visualization

Abstract

Stuttering affects about one percent of the world's population. Speech therapy for stutterers involves extended interaction between the patient and a skilled speech therapist. Computer aided speech and language therapy (CASLT) help provide therapy for more patients. CASLT has two main phases assignments of therapy to the stutterer and evaluation of progress. Current research has yet to address the progress evaluation phase. Therefore this research aims to develop an enhanced CASLT which can help user visualize his/her therapy progress. In this paper, we describe a computer aided Malay speech enhancer: a system that visualizes the speech therapy progress of stuttering Malay users. This system includes three main components. The Mel frequency cepstral coefficients, Vector quantization, and Hidden Markov model techniques are used to process the system’s audio signal. The student’s progress is visualized for him via extraction of his/her audio fingerprint.