Efficient Speaker Verification Algorithm using Spectral Characteristics
Vyshali V. Nayak
Vyshali V Nayak, B.E in E n C, Signal Processing, Cambridge Institute of Technology, VTU, Bangalore, Karnataka, India.
Manuscript received on 09 April 2016 | Revised Manuscript received on 16 April 2016 | Manuscript Published on 30 April 2016 | PP: 20-26 | Volume-5 Issue-11, April 2016 | Retrieval Number: K22840451116/2016©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Speaker recognition is a recognition purpose that articulates words. The speaker recognition process relies on physical structure of an individual’s person’s vocal tract and the behavioral characteristics of the individual. Speaker verification is evolved with the technologies of speech recognition and speech synthesis because of the similar characteristics in the voice and challenges associated with it. Speaker recognition has two forms which is text dependent or text independent. In text dependent method a particular phrase or password is stored into the system, whereas in text independent method the speaker will not be aware that his voice is being collected. In the proposed algorithm, speech signal has been recorded in the database. And the speaker is verified using the input the speaker provides by comparing with the database. The time domain, frequency domain and power domain features of the speech is extracted. For validating the performance, a comparative analysis has been carried out with various other methods. These methods exhibit some unique behavior.
Keywords: Spectral Characteristics, Speech Recognition, Text Dependent, Text Independent
Scope of the Article: Speech Recognition