الفهرس | Only 14 pages are availabe for public view |
Abstract The pronunciation (or phonetic) dictionary is an important ASR component that serves as an intermediate between language models and acoustic models in ASR systems. It holds a subset of the words available in the language and the pronunciation variants of each word in terms of sequences of the phonemes existing in the acoustic models. The development of perfect Automatic Arabic Speech Recognition (AASR) systems is challenged with the accuracy of the phonetic dictionary. Arabic is one of the phonetically complex languages and the creation of accurate speech recognition system is a challengeable task. The pronunciation variations in Arabic are tangible and are investigated widely using data driven approach or knowledge based approach. The phonological rules are used to get the pronunciation of each word accurately to reduce the mismatch between the actual phoneme representation of the spoken words and ASR dictionary. In this thesis we focus on those rules that handle within-word pronunciation variation and cross-word pronunciation variation in addition to building an acoustic model and language model that the decoder will use them to find the best sequence of words that match the input speech. The experimental results indicate that handling within-word pronunciation variation using phonological rule doesn’t enhance the recognition performance, but using these rules to handle cross-word variation provide a good performance in which the Word Error Rate(WER) is 10.47% compared with the base line system that has WER 11.27%. |