The Recognition of Holy Quran Reading types “Rewaih”

Abstract

This paper investigates recognition of a new trend in the holly Quran research it is Rewaih recognition. A data set of 78 verses (Aiah) was collected for a one reciter for two reading types Hfs Rewaih and Dori Rewaih. The system was trained using 38 verses, a separate set of 40 verses has been used for testing. For each Rewaih an HMM model has been trained. The test recognition rates for both Rewaih were 100%.

Keywords: The Holy Quran; Automatic speech recognition; Cepstral feature extraction; mixtures of Gaussian; Hidden Markov model .