Title :
Jacobian adaptation of noisy speech models
Author :
Sagayama, S. ; Yamaguchi, Y. ; Takahashi, S.
Author_Institution :
NTT Human Interface Labs., Tokyo, Japan
Abstract :
A Jacobian approach to fast adaptation of acoustic models is described. Acoustic models of speech under assumed noise and channel condition A are compensated by Jacobian matrices with the difference between condition A and actual condition B. Compared with existing model composition approaches for noisy speech recognition, this approach drastically reduces the computational cost while providing equivalent recognition performance. Extension of this analytic approach to acoustic model adaptation is also extensively discussed
Keywords :
Jacobian matrices; noise; performance evaluation; speech recognition; vectors; Jacobian adaptation; Jacobian matrices; acoustic model adaptation; channel condition; compensation; computational cost; model composition approaches; noise; noisy speech models; noisy speech recognition; performance; vectors; Acoustic noise; Adaptation model; Cepstrum; Jacobian matrices; Loudspeakers; Noise reduction; Signal to noise ratio; Speech enhancement; Speech recognition; Working environment noise;
Conference_Titel :
Automatic Speech Recognition and Understanding, 1997. Proceedings., 1997 IEEE Workshop on
Conference_Location :
Santa Barbara, CA
Print_ISBN :
0-7803-3698-4
DOI :
10.1109/ASRU.1997.659116