Speech rehabilitation for disorder people

Ahmad Azhar, Nur Syahmina (2022) Speech rehabilitation for disorder people. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Preview

Text (Full Text) Speech rehabilitation for disorder people.pdf - Submitted Version Download (3MB) | Preview

Abstract

Communication is one of the crucial elements in life. It helps humans make decisions, solve problems, share ideas, and develop relationships. Effective communication will help people in many aspects, especially in the medical field. However, this effective communication is a limitation for some people who have problems communicating, like stroke patients, children with speech delay, Aphasia patients, and more. For gaining their capability in speech, rehabilitation is one method to retrain back this lost capability. Here, the major problems for the speech capability among the patient have commonly affected patients’ understanding of speeches and have disorganized speech. Although manual rehabilitation is the most common way to train them, there are still problems of housekeeping and human power in ensuring the training and rehabilitation activities are on track. Lack of human strength and the effect of pandemics infers the delay time of patient rehabilitation programs. This project showed an outperform vowel recognition result via Convolution Neural Network (CNN) via vowel sound dataset. Compared to the comparative and the baseline method, VGG16, we manage to recognize the vowel using not directly sound but from an image file. Our proposed work could contribute to the medical and rehabilitation field helping the disorder people by gaining this outperform accuracy. At the end of this project, a network model, which will be utilized as Malay Language vowel detection with good performance accuracy, can be achieved.

Actions (login required)

View Item