N K Narayanan¹ T M Thasleema¹ and V Kabeer², ¹Kannur University, India and ²University of Calicut, India

In this paper we extend Support Vector Machines (SVM) for speaker independent Consonant – Vowel (CV) unit classification. Here we adopt the technique known as Decision Directed Acyclic Graph (DDAG) , which is used to combine many two class classifiers into multiclass classifier. Using Reconstructed State Space (RSS) based State Space Point Distribution (SSPD) parameters, we obtain an average speaker independent phoneme recognition accuracy of 90% on the Malayalam V/CV speech unit database. The recognition results indicate that this method is efficient and can be adopted for developing a complete speech recognition system for Malayalam language.

Reconstructed State Space, State Space Map, State Space Point Distribution Parameter, Support Vector Machine