Detection of Mad Lazim Harfi Musyba Images Uses Convolutional Neural Network

Abstract

Reading of Al Quran is an obligation for all Muslims. Lack of knowledge about the knowledge of recitation in reading the Al Quran is certainly a problem. The purpose of this study makes it easier for everyone to learn the law of recitation, especially Madd Lazim Harfi Musyba in the Al Quran verses. The method used is one of the deep neural network methods, namely Convolutional Neural Network (CNN), as real-time detection of the law of Madd Lazim Harfi Musyba, implementation of the method using the help of Tensorflow GPU (Graphic Processor Unit) library. The results of trials with the Deep Convolutional Neural Network model show the detection performance of 9 verses with an average accuracy of 93.25%. The conclusion is that the training data on the CNN model is very reliable in detecting the Mad lazim harfi musyba law. Therefore this system can be used to assist in applying the Mad Lazim Hafi Musyba legal recitation while reading the Al Quran.