Multi-stage speaker adaptation

Patent metadata

Field	Value
Publication number	US-8996366-B2
Application number	US-201414181908-A
Country	US
Kind code	B2
Filing date	Feb 17, 2014
Priority date	May 31, 2012
Publication date	Mar 31, 2015
Grant date	Mar 31, 2015

Abstract

A first gender-specific speaker adaptation technique may be selected based on characteristics of a first set of feature vectors that correspond to a first unit of input speech. The first set of feature vectors may be configured for use in automatic speech recognition (ASR) of the first unit of input speech. A second set of feature vectors, which correspond to a second unit of input speech, may be modified based on the first gender-specific speaker adaptation technique. The modified second set of feature vectors may be configured for use in ASR of the second unit of input speech. A first speaker-dependent speaker adaptation technique may be selected based on characteristics of the second set of feature vectors. A third set of feature vectors, which correspond to a third unit of speech, may be modified based on the first speaker-dependent speaker adaptation technique.

First claim

What is claimed is:

1. A method comprising:
based on characteristics of a first unit of input speech, selecting a gender-specific speaker adaptation technique, wherein the gender-specific speaker adaptation technique relates to a particular gender;
conducting, by a computing device, automatic speech recognition (ASR) of a second unit of input speech, wherein the ASR of the second unit of input speech is based on the gender-specific speaker adaptation technique;
while conductin…

Assignees

• Google Inc

Inventors

• Aleksic Petar
• Lei Xin

Classifications

• G10L15/07 Primary

  Physics · mapped topic

• G10L15/065

  Physics · mapped topic

• G10L17/00 Primary

  Physics · mapped topic