Research

Ongoing Projects

MULTIMODAL ENGAGEMENT FOR CHILDREN WITH COMMUNICATION DISABILITIES (UM.C/HIR/MOHE/FCSIT/05)
This project will contribute to the Malaysian Disabled Society especially for children with communication disabilities in improving the quality of their life and to be more independent. With improvement to their pronunciation, it catalyzed the improvement of verbal communication among the Malay speaking speech impaired children with people close to them such as parents, caregivers and teachers.

Project Members

Project1

(Source: Mumtaz et al., 2014 Severity-based adaptation with limited  data for ASR to aid dysarthric speakers)

DYSARTHRIC SPEECH RECOGNITION USING MULTI-VIEWS MULTI-LEARNERS (UM.C/HIR/MOHE/FCSIT/05)
Automatic Speech Recognition (ASR) can be very helpful for speakers who suffer from dysarthria, a neurological disability that damages the control of motor speech articulators. Although a few attempts have been made to apply ASR technologies to sufferers of dysarthria, previous studies show that such ASR systems have not attained an adequate level of performance. In this study, a speech recognition model for dysarthric speakers is studied using a realisation of Multi-Views Multi-Learners approach called Multi-Nets Artificial Neural Networks. The objectives of this study are 1) to investigate which speech features are more robust in order to represent dysarthria, 2) to improve the accuracy of ASR, and 3) to propose a speaker-independent dysarthric ASR model.

Project Members

Project2

(Source: Seyed Reza Shahamiri et al., 2014 Identifying the best-performing set of MFCC parameters and studying a speaker-independent approach, Advanced Engineering Informatics, Elsevier)

AN AUTOMATIC SPEECH PRONUNCIATION PRACTICE AND FEEDBACK SYSTEM FOR CHILDREN WITH SPEECH IMPAIRMENTS (RP002E-13ICT)
In this research, we aim to develop the state-of-the-art speech pronunciation practice system to provide automatic feedback and web-based evaluation of speech-impaired children’s pronunciation. This form of feedback system can made publicly available on the web, where many children will have access to the feedback system anywhere and anytime with minimal involvement of speech therapist.

Project Members

Project3

(Source: Azila Ahmad, Automatic speech recognition in pronunciation practice game application for speech impaired children. Master Thesis 2012.)

Completed Projects

HMM-BASED SPEECH SYNTHESIS SYSTEM FOR MALAY
The-state-of-the-art speech synthesis systems such as the one based on Hidden Markov Model (HMM) provide an opportunity to synthesize speech with acceptable naturalness. However, such development requires adequate resources including segmented and labelled of speech data, which is a problem for many under-resourced languages, including Malay. This research shows how it is possible to build a speech acoustic model for Malay with minimal resources.

Project Members

Project4

(Source: Mumtaz et al., 2014 , IEICE Transactions on Information and System)