US Patent 7454338 - Training wideband acoustic models in the cepstral domain using mixed-bandwidth training data and extended vectors for speech recognition

Training wideband acoustic models in the cepstral domain using mixed-bandwidth training data and extended vectors for speech recognition

Document Number

US Patent 7454338

Issued Date

November 18, 2008

Link

Inventors

Seltzer; Michael L. (Seattle, WA)

show on map

Acero; Alejandro (Bellevue, WA)

show on map

Map

Abstract

A method and apparatus are provided that generate values for a first set of dimensions of a feature vector from a speech signal. The values of the first set of dimensions are used to estimate values for a second set of dimensions of the feature vector to form an extended feature vector. The extended feature vector is then used to train an acoustic model.