Volume 11 Issue 2 ( June )

Voice identification is the process to identify a human, based on his/her voice. In this paper, we propose a voice identification system using a technology of 5-Means Clustering approach on both Bandwidth and Formants Ratio. At first, the PRAA database is trained by these methods, this method has used five clustered for bandwidth and also formants ratio. These clusters are classified on the behalf of fuzzy members, and thereafter each and every cluster of bandwidth and formants ratio are combined in different type of combinations. In the testing section, voice samples taken from speakers are compared with these combinations. These systems make them provided fast voice identification process. The Experiment result of the proposed work is achieved more than 98% of Identification Rate.

Keywords: Bandwidth, Clustering, Feature Extraction, Formants, Fuzzy, Voice Identification.