Combination of Vector Quantization and Hidden Markov Models for Arabic Speech Recognition

ACS/IEEE International Conference on Computer Systems and Applications (AICCSA'01) p. 0096
Combination of Vector Quantization and Hidden Markov Models for Arabic Speech Recognition
H. Bahi, University of Annaba
M. Sellami, University of Annaba

Full Article Text:

DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/AICCSA.2001.933957
Send link to a friend

Abstract
Abstract: In this paper we present experiments we perform in order to recognize Arabic isolated words. Our recognition system is based on the combination of the vector quantization technique at the acoustical level and the Markovian modelling. The Hidden Markov Models (HMMs) are widely used in number of practical applications and especially suitable in speech recognition because of their ability to handle the variability of the speech signal. In our system, a word is analysed and represented as a set of acoustical vectors then transformed into a symbolic sequence, using the vector quantizer. This observation sequence is compared to the reference Markov models. The word associated to the model who obtained the highest score is declared the recognized word.
Back to Top Additional Information
Index Terms- speech recognition, vector quantization, hidden Markov models, artificial intelligence, training. Citation: H. Bahi, M. Sellami, "Combination of Vector Quantization and Hidden Markov Models for Arabic Speech Recognition," aiccsa, p. 0096, ACS/IEEE International Conference on Computer Systems and Applications (AICCSA'01), 2001

	Advanced Search
CS Search Google Search