No Detail Found!

Automatic speech recognition on non-pathological dataset of urdu language

Author(s)
Anoshia Imtiaz Biomedical Engineering Department, Ziauddin University Faculty of Engineering Science Technology and Management, Karachi, Pakistan
Hira Zahid Electrical Engineering Department & Software Engineering Department, Ziauddin University Faculty of Engineering Science Technology and Management, Karachi, Pakistan
Munaf Rashid Biomedical Engineering Department, Sir Syed University of Engineering & Technology, Karachi, Pakistan
Muzzaffar Iqbal Biomedical Engineering Department, Ziauddin University Faculty of Engineering Science Technology and Management, Karachi, Pakistan
Sidra Abid Syed Biomedical Engineering Department, Ziauddin University Faculty of Engineering Science Technology and Management, Karachi, Pakistan
Akhtar Ali Khan Biomedical Engineering Department, Ziauddin University Faculty of Engineering Science Technology and Management, Karachi, Pakistan
Abstract
Voice is a primary tool for communications and voice disorders bring atypical characteristics in the voice which influence the quality of voice. Voice disorders are abnormal conditions that influence the quality of voice. Several protocols, including acoustic analysis, can detect clinical voice pathology. Based on the computerized acoustic analysis, machine learning algorithms and non-invasive systems may play a vital part in the initial detection, tracking, and growth of proficient pathological speech analysis. The methodology proposes to collect a non-pathological dataset, i.e., a healthy voice dataset, and offers a unique combination of feature extraction techniques combining MelFrequency Cepstral Coefficients (MFCC), and Pitch. Support Vector Machine (SVM) was used as a machine learning classifier for the training and testing of the dataset model. The SVM algorithms demonstrated satisfactory training and testing accuracy rate, i.e., 85.886%, which proves to be a milestone on the Urdu language dataset.
Publication Details
Page(s) 117-126
Published Journal: KIET Journal of Computing & Information Sciences, Volume: 5, Issue: 2, Year: 2022

Request the full text from PASTIC

Keywords
SVM, Urdu Language, MFCC, Voice dataset, Pitch.
0

Citations

1

Downloads

70

Views