Identifying the correct articulation point of a Quranic letters of the throat (al-halqu) makhraj

Othman, Ahmad Al Baqir and Ahmad, Salmiah and Badron, Khairayu and Altalmas, Tareq M. K. (2023) Identifying the correct articulation point of a Quranic letters of the throat (al-halqu) makhraj. PERINTIS eJournal, 13 (2). pp. 40-64. ISSN 2232-0725

PDF - Published Version Restricted to Registered users only Download (1MB) | Request a copy

Abstract

Tajweed is an essential tool for a Muslim in reciting the Holy Quran as it ensures the recitation is properly done as what has been practiced by the Prophet Muhammad s.a.w and his companions. In Tajweed, mastering makhraj, or the Quranic letter point of articulation in pronunciation, is crucial for a Muslim. Tajweed is traditionally taught in "Talaqqi", where face�to-face sessions are commonly conducted with teachers, from early ages children. The fact is that the time and resources are limited especially in some areas in the world, where getting easy access towards a certified teacher for Quranic teaching and learning is a big issue. Considering this issue, with the enrichment of technology particularly in Artificial Intelligence (AI) nowadays, this limitation can be overcome. The conventional face-to-face learning can be supported with AI-based system for Quranic teaching and learning. There are few systems with regard to Quranic teaching and learning but most of them are focusing on memorization and recitation of the Quran rather than the Tajweed. This study presents the algorithm design, technique, and simulation of a Speech Recognition-based method to detect the correct articulation point (Makhraj) of the throat letters in the Quranic word. The study starts with a lots of data collection from experts in pronouncing the Quranic letters via developed platform and Apps called SanaAlTajweed, data augmentation, data pre-processing, features extraction using MFCC and deep learning model and validation using designed Graphical User Interface (GUI). Data was trained using an improved deep learning Convolutional Neural Network (CNN) classification model. Results shows that the algorithm was able to detect the letters produced at throat area, which are also known as Izhar Halqi letters. The rresults obtained have shown a promising achievement where the model has attained an accuracy of 89.39% on the test dataset with a loss of 0.4305. The outcomes may help to provide the most efficient way for machines to analyse the Quranic recitation and translate it into a new model that will help human and computer to understand, interact, and improve.

Item Type:	Article (Journal)
Uncontrolled Keywords:	MFCC, CNN, Data Augmentation, Makhraj, Tajweed, Quran
Subjects:	T Technology > T Technology (General) > T173.5 Technology and Islam
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Engineering > Department of Mechanical Engineering Kulliyyah of Engineering Kulliyyah of Engineering > Department of Electrical and Computer Engineering
Depositing User:	Ir. Dr. Salmiah Ahmad
Date Deposited:	11 Jun 2024 15:05
Last Modified:	11 Jun 2024 15:05
URI:	http://irep.iium.edu.my/id/eprint/112593

Actions (login required)

View Item