 |
ITRW on
|
 |
|
ITRW on
|
|
We present a fast method, i.e. requiring little data, for adapting a hybrid Hidden Markov Model / Multi Layer Perceptron speech recognizer to reverberant environments. Adaptation is performed by a linear transformation of the acoustic feature space. A dimensionality reduction technique similar to the eigenvoice approach is also investigated. A pool of adaptation transformations are estimated a priori for various reverberant environments. Then, the principal directions of the pool are extracted, the so-called eigenrooms. The adaptation transformation for every new reverberant environment is constrained to lay on the subspace spanned by the most significant eigenrooms. Consequently, the adaptation procedure involves estimating only the projection coefficients on the selected eigenrooms, which requires less data than direct estimation of the adaptation transformation. Supervised adaptation experiments for recognition of connected digit sequences (AURORA database) in reverberant environments are carried out. Standard adaptation demonstrates improvements in word error rate higher than 30% for typical reverberation levels. The eigenroom-based adaptation technique implemented so far allows at most 50% reduction of adaptation data for the same improvement.
Bibliographic reference. Couvreur, Laurent / Dupont, S. / Ris, C. / Boite, J.-M. / Couvreur, Christophe (2001): "Fast adaptation for robust speech recognition in reverberant environments", In Adaptation-2001, 85-88.