Towards using CMU sphinx tools for the Holy Quran recitation verification

The use of the Automatic Speech Recognition (ASR) technology is being used in many different applications that help simplify the interaction with a wider range of devices. This paper investigates the use of a simplified set of phonemes in an ASR system applied to Holy Quran. The Carnegie Mellon University Sphinx 4 tools were used to train and evaluate a language model on Holy Quran recitations that are widely available online. The building of the language model was done using a simplified list of phonemes instead of the mainly used Romanized in order to simplify the process of training the acoustic model. In this paper, the experiments resulted in Word Error Rates (WER) as low as 1.5% even with a very small set of audio files used during the training phase