R. Thiruvengatanadhan (Assistant Professor)
Annamalai University, Annamalainagar, Tamil Nadu, India.
Automatic recognition of speech to make fast between human and machine
communication. This paper describes a technique that uses Support Vector Machine (SVM)
and Gaussian Mixture Model (GMM) to recognized speech based on features using Power
Normalized Cepstral Coefficients (PNCC). Displaying methods, for example, SVM and GMM
were utilized to demonstrate every individual word which is prepared to the framework. Each
segregated word Segment utilizing Voice Activity Detection (VAD) from the test sentence is
coordinated against these models for finding the semantic portrayal of the test input
discourse. Experimental results of GMM shows good performance in recognized rate.