Text Independent Speech Recognition

Voice is one of the well-researched biometrics for speaker authentication. So, in the literature a number of techniques have been proposed which make use of speaker’s voice to verify their identity, control access to services, utilize information services, voice mailing, security control for confidential information areas, and remote access to computers. Speaker recognition is the task of automatically recognizing the speaker among several reference speakers using speaker-specific information embedded in speech waves. The extraction of this information is termed as features, and matching process are implemented right after the pre-processing of the signal. In literature, techniques have been proposed where Mel-Frequency Cepstral Coefficients (MFCCs) are taken as features for modeling by the Gaussian Mixture Model (GMM) during the identification process. The GMM algorithm is one of the clustering algorithms employed for the text independent speaker recognition. In this context, by adopting the idea of hierarchical clustering and GMM, an adaptive clustering algorithm, i.e., AGMM that can determine the number of optimal clusters automatically has been proposed. Numerical experiments demonstrate that the AGMM achieves better performance than the traditional GMM where the number of clusters is fixed.