Comparative Study of ANN and HMM to Arabic Digits Recognition Systems

Abstract

Arabic language is a Semitic language that has many differences when compared to Latin languages such as English. One of these differences is how to pronounce the ten digits, zero through nine. All Arabic digits are polysyllabic (except digit zero which is a monosyllabic) words and most of them contain Arabic unique phonemes, namely, pharyngeal and emphatic subset. In a previous paper the researcher designed an Artificial Neural Networks (ANN) based Arabic digits recognition system. In this paper we continued the research by designing Hidden Markov Model (HMM) based system that was designed and tested with automatic Arabic digits recognition. The old system was isolated whole word speech recognizer, but the current one was an isolated word phoneme based recognizer. Both systems were implemented both as a multi-speaker (i.e., the same set of speakers were used in both the training and testing phases) mode and speaker-independent (i.e., speakers used for training are different from those used for testing) mode. The main aim of this paper was to compare, analyze, and discuss the outcomes of these two recognition systems. The ANN based recognition system achieved 99.5% correct digit recognition in the case of multi-speaker mode, and 94.5% in the case of speaker-independent mode. On the other hand, the HMM based recognition system achieved 98.1% correct digit recognition in the case of multi-speaker mode, and 94.8% in the case of speaker-independent mode.