A Classifier Combination Approach for Farsi Accents Recognition
Keywords:
automatic speech recognition, accentclassification, phonotactic approach, acoustic approach,classifier combination
Abstract:
Accent classification technologies directly influencethe performance of automatic speech recognition (ASR)systems. In this paper, we evaluate three accent classificationapproaches: Phone Recognition followed by LanguageModeling (PRLM) as a phonotactic approach; accent modelingusing Gaussian Mixture Models (GMM) then selecting the mostsimilar model using Maximum Likelihood algorithm that iscategorized in acoustic approaches a novel classifiercombination method which is proposed to improve theperformance of accent classification for several regionalaccents. In the proposed approach, we use an ensemble methodin which each base classifier is a binary classifier thatseparates an accent from another one. We use the majority votealgorithm to combine the base classifiers. Results for fiveaccents selected from FARSDAT speech database show that theproposed ensemble method outperforms PRLM and GMMbasedapproaches in the case of Farsi regional accentclassifications.