Holy Qur'an speech recognition system Imaalah checking rule for warsh recitation

Abstract

In the process of learning Quran, reciters should have the provisions of Tajweed rules when reading the Quran. This study provides a research related to Quran that can assist and ensure proper pronunciation, readings and interpretations in proper Quranic recitation rules based on the speech recognition where reciters can distinguish and recognize and correct during their recitation the pronunciation of Imaalah (Tajweed rules) for Warsh recitation type. In this research, speech samples form reciters is used to compare with the sample recitation features collected from expert reciters that have been stored in the database. The collected samples are analysed by using preprocessing techniques. The sound features are extracted during the features extraction using the well-known algorithm which is Mel-Frequency Cepstral Coefficient (MFCC). Subsequently, after features extraction, a classification algorithm (Hidden Markov Models (HMM)) is employed to compare the features extracted in real-time with that available in the knowledge base.