VALIDATION D’UNE RÉCITATION CORANIQUE EN UTILISANT UN CNN

Abstract

The Quran is considered the primary source of Islamic law, but many find it difficult to memorize and recite it accurately. Deep learning models have emerged as an effective means to facilitate this process. In this work, we introduced a novel deep learning approach using a CNN model to identify the numbers of verses and sourah from audio recordings of Quranic recitation, thus facilitating memorization. We trained and tested the model using the "quran reciters" database, which proved effective with an accuracy rate of 95% on the test data.