Automatic Classification of Standard Arabic Phonemes Using Parallel Genetic Algorithms

Abstract

The aim of this work is the application of the Parallel Genetic Algorithms (PGAs) to the Automatic Speech Recognition (ASR) domain at the acoustic sequences classification level. Speech recognition was been cast as a pattern classification problem where we would like to classify an input acoustic signal into one of all possible phonemes. Thus, we have looked for recognizing Standard Arabic (SA) phonemes of continuous, naturally spoken, speech by using PGAs whose have several advantages in resolving complicated optimization problems. The phonemes classification duration or problem resolution time is reduced by the number of the Genetic Algorithms (Gas) used in parallel, simultaneously. In SA, there are forty sounds. We have analysed a corpus that contains several sentences composed of the whole SA phonemes types in the initial, medium and final positions, recorded by several male speakers, in low noisy environment. Thus, the acoustic segments classification and the PGAs have been explored. We have used the decision rule Manhattan distance as the fitness functions for the PGAs evaluations whose topology chosen is the isolated island one. The Corpus phonemes were extracted and classified successfully with an overall accuracy of  95, 75% more rapidly than in the classical GAs methods. In addition, the computational cost was greatly reduced, and so the performances of the GAs were improved.