A Classifier Combination Approach for Farsi Accents Recognition

Accent classification technologies directly influence the performance of automatic speech recognition (ASR) systems. In this paper, we evaluate three accent classificationapproaches: Phone Recognition followed by Language Modeling (PRLM) as a phonotactic approach; accent modeling using Gaussian Mixture Models (GMM) then selecting the mostsimilar model using Maximum Likelihood algorithm that is categorized in acoustic approaches a novel classifiercombination method which is proposed to improve the performance of accent classification for several regional accents. In the proposed approach, we use an ensemble methodin which each base classifier is a binary classifier that separates an accent from another one. We use the majority votealgorithm to combine the base classifiers. Results for five accents selected from FARSDAT speech database show that the proposed ensemble method outperforms PRLM and GMMbased approaches in the case of Farsi regional accent classifications.

Keywords:

automatic speech recognition , accent classification , phonotactic approach , acoustic approach , classifier combination