TY - GEN
T1 - Development of teo phase for speaker recognition
AU - Patil, Hemant A.
AU - Parhi, Keshab K.
PY - 2010/10/29
Y1 - 2010/10/29
N2 - Most of the speaker recognition systems use system features for speaker recognition which are mostly spectral in nature. Recently, there has been significant work on using source features, viz., prosodies and pitch dynamics, glottal flow derivative, Linear Prediction (LP) residual and its phase, wavelet-domain representation of LP residual, etc for speaker recognition. In this paper, a new source-like feature set, viz., Teager Energy Operator (TEO) Phase is developed for speaker recognition. Proposed TEO Phase has several salient features as compared to that of LP residual phase, viz., less number of spurious peaks in Hilbert envelope plot, and better perceptual and sample correlation with speech signal. Furthermore, this avoids use of voiced/unvoiced detection, preprocessing techniques such as windowing and pre-emphasis and LP residual computation. Experiments have been carried out for speaker recognition task using TEO Phase and LP residual phase with polynomial classifier of 2 nd order approximation. For speaker verification, a reduction in equal error rate (EER) by 2.62 % over LP residual phase is achieved for proposed feature set.
AB - Most of the speaker recognition systems use system features for speaker recognition which are mostly spectral in nature. Recently, there has been significant work on using source features, viz., prosodies and pitch dynamics, glottal flow derivative, Linear Prediction (LP) residual and its phase, wavelet-domain representation of LP residual, etc for speaker recognition. In this paper, a new source-like feature set, viz., Teager Energy Operator (TEO) Phase is developed for speaker recognition. Proposed TEO Phase has several salient features as compared to that of LP residual phase, viz., less number of spurious peaks in Hilbert envelope plot, and better perceptual and sample correlation with speech signal. Furthermore, this avoids use of voiced/unvoiced detection, preprocessing techniques such as windowing and pre-emphasis and LP residual computation. Experiments have been carried out for speaker recognition task using TEO Phase and LP residual phase with polynomial classifier of 2 nd order approximation. For speaker verification, a reduction in equal error rate (EER) by 2.62 % over LP residual phase is achieved for proposed feature set.
KW - LP residual phase
KW - Speaker recognition
KW - TEO Phase
UR - http://www.scopus.com/inward/record.url?scp=77958461504&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77958461504&partnerID=8YFLogxK
U2 - 10.1109/SPCOM.2010.5560486
DO - 10.1109/SPCOM.2010.5560486
M3 - Conference contribution
AN - SCOPUS:77958461504
SN - 9781424471362
T3 - 2010 International Conference on Signal Processing and Communications, SPCOM 2010
BT - 2010 International Conference on Signal Processing and Communications, SPCOM 2010
T2 - 2010 International Conference on Signal Processing and Communications, SPCOM 2010
Y2 - 18 July 2010 through 21 July 2010
ER -