Evaluation of the Pronunciation of Tajweed Rules Based on DNN as a Step Towards Interactive Recitation Learning

2025-03-30

AI Summary

To address the limitations of manual Tajweed rule assessment in Quranic recitation—namely, scarcity of qualified instructors and low evaluation efficiency—this paper proposes a deep learning–based automatic pronunciation assessment method targeting three core Tajweed rules: Al Mad, Ghunnah, and Ikhfaa. We introduce a novel end-to-end speech classification model that integrates EfficientNet-B0 with the Squeeze-and-Excitation channel attention mechanism. Mel-spectrogram features coupled with normalization-based preprocessing are employed. Evaluated on the public QDAT dataset, the model achieves fine-grained recognition accuracies of 95.35%, 99.34%, and 97.01% for the respective rules. The proposed approach demonstrates high accuracy, strong robustness to acoustic variability, and superior generalization across diverse reciters and recording conditions. It enables instructor-free, self-paced practice with real-time feedback, offering a scalable, deployable technical framework for intelligent religious language education.

Technology Category

• Machine Learning: Evaluation and Analysis
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Proper recitation of the Quran, adhering to the rules of Tajweed, is crucial for preventing mistakes during recitation and requires significant effort to master. Traditional methods of teaching these rules are limited by the availability of qualified instructors and time constraints. Automatic evaluation of recitation can address these challenges by providing prompt feedback and supporting independent practice. This study focuses on developing a deep learning model to classify three Tajweed rules - separate stretching (Al Mad), tight noon (Ghunnah), and hide (Ikhfaa) - using the publicly available QDAT dataset, which contains over 1,500 audio recordings. The input data consisted of audio recordings from this dataset, transformed into normalized mel-spectrograms. For classification, the EfficientNet-B0 architecture was used, enhanced with a Squeeze-and-Excitation attention mechanism. The developed model achieved accuracy rates of 95.35%, 99.34%, and 97.01% for the respective rules. An analysis of the learning curves confirmed the model's robustness and absence of overfitting. The proposed approach demonstrates high efficiency and paves the way for developing interactive educational systems for Tajweed study.

Problem

• Develops DNN model to classify Tajweed rules for Quran recitation
• Addresses lack of instructors with automated pronunciation evaluation
• Uses mel-spectrograms and EfficientNet-B0 for high-accuracy rule detection

Innovation

• Uses DNN for Tajweed pronunciation evaluation
• Applies EfficientNet-B0 with attention mechanism
• Achieves high accuracy on QDAT dataset