Klasifikasi Sifat Huruf Hijaiyah Dengan Metode Convolutional Neural Network (CNN)

Abstract

Understanding the science of Tajweed, particularly the articulatory properties of hijaiyah letters, plays a crucial role in enchancing the quality of Quranic recitation. Despite its importance, research focused on classifying these properties within Quranic texts remains limited. Existing Tajweed learning tools often introduce letters at a basic level without utilizing deep learning technologies. This study proposes a CNN-based model to classify the phonetic characteristics of hijaiyah letters in Quranic texts. The dataset consists of image samples taken from quran.com, each labeled according to the phonetic categories outlined in the Tartil Al-Quran guidebook. The methodology includes image preprocessing, CNN training, and performance evaluation using accuracy, precision, recall, and F1-score. This research does not address audio or pronunciation aspects. Results show that the model achieved up to 99% classification accuracy. The findings highlight the potential of AI-powered tools to support Tajweed learning and contribute to the development of intelligent, technology-based Quranic education systems. This research serves as a foundation for future applications that blend classical Islamic knowledge with modern digital solutions.