A novel approach for automatic speaker identification employing a recently proposed signal compression technique, namely, the multiple transform domain split vector quantiser, is developed and presented. Using a normalised matching accuracy measure, the proposed technique consistently yields enhanced identification performance in comparison with existing single domain vector quantiser approaches.
References
-
-
1)
-
W.B. Mikhael ,
V. Krishnan
.
Multiple transform domain split vector quantisation.
Electron. Lett.
,
538 -
539
-
2)
-
A. Gray ,
R.M. Gray
.
(1991)
Vector quantisation and signal compression.
-
3)
-
Tufekci, Z., Gowdy, J.N.: `Feature extraction using discrete wavelet transform for speech recognition', IEEE Proceedings, April 2000, Southeastcon, Nashville, TN, USA, p. 116–123.
-
4)
-
Berg, A.P., Mikhael, W.B.: `A survey of mixed transform techniques for speech and image coding', IEEE, Int. Symp. on Circuits and Systems, June 1999, Orlando, FL, USA, p. 106–109.
http://iet.metastore.ingenta.com/content/journals/10.1049/el_20020948
Related content
content/journals/10.1049/el_20020948
pub_keyword,iet_inspecKeyword,pub_concept
6
6