Dynamic Bayesian Networks for Multi-Dialect Isolated Arabic Recognition

Hidden Markov Models (HMM) are currently widely used in Automatic Speech Recognition (ASR) as being the most effective models. In addition, the HMM are just a special case of graphical models which are dynamic Bayesian Networks (DBN). These are modeling tools more sophisticated because they allow to include several specific variables in the problem of automatic speech recognition other than the one used in HMM. The use of DBNs in speech recognition beyond has generated much interest in recent years. This paper deals a comparative study between DBN and HMM systems for multi-dialect isolated Arabic words. Performance using DBNs is found to exceed that of HMMs trained on an identical task, giving higher recognition accuracy for four different Arabic dialects. In fact, the average of recognition rates for the four dialects with DBN was 83.01% while 74.17% with HMM standards.