Speaker adaptation of vocabulary for speech recognition

• US 20040117180A1
• Filed: 12/16/2002
• Published: 06/17/2004
• Est. Priority Date: 12/16/2002
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Rajput, Nitendra, Verma, Ashish
• Granted Patent: US 7,389,228 B2
• Time in Patent Office: Days
• US Class Current: 704/231
• CPC Class Codes: G10L 15/07 to the speaker
• CPC Class Codes: G10L 17/02 Preprocessing operations, e.g. segment selection; Pattern representation or modelling, e.g. based on linear discriminant analysis [LDA] or principal components; Feature selection or extraction

Abstract

A phonetic vocabulary for a speech recognition system is adapted to a particular speaker'"'"'s pronunciation. A speaker can be attributed specific pronunciation styles, which can be identified from specific pronunciation examples. Consequently, a phonetic vocabulary can be reduced in size, which can improve recognition accuracy and recognition speed.

Claims

1. A method for speaker adaptation of vocabulary for a speech recognition system, the method comprising the steps of:

attributing a pronunciation style to the speaker, based upon the identified pronunciation; and

restricting a vocabulary used in the speech recognition system to pronunciations consistent with the attributed pronunciation style.

2. The method as claimed in claim 1, further comprising the step of:

categorizing words in a vocabulary in accordance with patterns in alternative word pronunciations.

3. The method as claimed in claim 1, further comprising the step of:

recording an enrollment text, including predetermined keywords, read by the speaker.

4. The method as claimed in claim 3, wherein the keywords are representative of words having a phonetically similar baseform.

5. The method as claimed in claim 4, wherein the keywords are representative of words having alternative pronunciations that are indicative of a speaker'"'"'s likely pronunciation of similarly categorized words.

6. The method as claimed in claim 5, wherein the step of attributing a pronunciation style is based upon the speaker'"'"'s pronunciation of the keywords recorded in the enrollment text.

7. The method as claimed in claim 1, wherein the step of attributing a pronunciation style is based upon the speaker'"'"'s pronunciation of predetermined keywords.

8. A computer system for speaker adaptation of vocabulary for a speech recognition system, the computer system comprising:

means for attributing a pronunciation style to the speaker, based upon the identified pronunciation; and

means for restricting a vocabulary used in the speech recognition system to pronunciations consistent with the attributed pronunciation style.

9. Computer software, recorded on a medium, for vocabulary adaptation for a speech recognition system, the computer software comprising:

software code means for attributing a pronunciation style to the speaker, based upon the identified pronunciation; and

software code means for restricting a vocabulary used in the speech recognition system to pronunciations consistent with the attributed pronunciation style.