For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Adaptation of Acoustic Models in Joint Speaker and Noise Space Using Bilinear Models
Yongwon JEONG Hyung Soon KIM
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2014/08/01
Online ISSN: 1745-1361
Type of Manuscript: LETTER
Category: Speech and Hearing
bilinear model, eigenvoice speaker adaptation, environment adaptation, speaker adaptation, speech recognition,
Full Text: PDF(100.3KB)>>
We present the adaptation of the acoustic models of hidden Markov models (HMMs) to the target speaker and noise environment using bilinear models. Acoustic models trained from various speakers and noise conditions are decomposed to build the bases that capture the interaction between the two factors. The model for the target speaker and noise is represented as a product of bases and two weight vectors. In experiments using the AURORA4 corpus, the bilinear model outperforms the linear model.