Isolated Word Recognition System for Autistic Speech

This paper aims to discuss the implementation of an Isolated Word Recognition system for Autistic speech using Hidden Markov Model (HMM). This system will be able to recognize the spoken words of Autistic children. Autism is a neurological disability which resists the functions of neurons in human brain. The most common problem with autistic children is communication. Focusing on their speech, it is observed that there is a high variability compared to normal speakers’ speech and it is unclear and not always understandable. Considering this, we aim to build a system to recognize the autistic speech and give the output as text and convert it back to clear speech using speech synthesizers.