Robust Features for Automatic Text-Independent Speaker Recognition using Gaussian Mixture Model
R. Rajeshwara Rao1, A. Prasad2, Ch. Kedari Rao3
1R. Rajeshwara Rao, Professor & Head, Department of Computer Science & Engineering, MGIT, Hyderabad, India.
2A. Prasad, Professor & Head, Department of Computer Applications, Vignan University, Guntur, India.
3Ch. Kedari Rao, Assistant Professor, Department of Computer Science & Engineering, DVRCET, Hyderabad, India.
Manuscript received on October 13, 2011. | Revised Manuscript received on October 27, 2011. | Manuscript published on November 05, 2011. | PP: 330-335 | Volume-1 Issue-5, November 2011. | Retrieval Number: E0239101511/2011©BEIESP
Open Access | Ethics and Policies | Cite
© The Authors. Published By: 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: In this paper, robust features for text-independent speaker recognition has been explored. Through different experimental studies, it is demonstrated that the speaker related information can be effectively captured using Gaussian mixture Models (GMMs). The study on the effect of feature vector size for good speaker recognition demonstrates that, feature vector size in the range of 20-24 can capture speaker discrimination information effectively for a speech signal sampled at 16 kHz, it is established that the proposed speaker recognition system requires significantly less amount of data during both during training as well as in testing. The speaker recognition study using robust features for different mixtures components, training and test duration has been exploited. We demonstrate the speaker recognition studies on TIMIT database.
Keywords: Gaussian Mixture Model ( GMM), MFCC, Robust Features, Speaker.