Compound words in speech recognition systems

Patent 6349282 Issued on February 19, 2002. Estimated Expiration Date:

April 20, 2019. Estimated Expiration Date is calculated based on simple USPTO term provisions. It does not account for terminal disclaimers, term adjustments, failure to pay maintenance fees, or other factors which might affect the term of a patent.

Abstract Claims Description Full Text

Patent References

Natural language understanding system
Patent #: 4914590
Issued on: 04/03/1990
Inventor: Loatman, et al.

Language translation apparatus and method using context-based translation models
Patent #: 5510981
Issued on: 04/23/1996
Inventor: Berger, et al.

Speech recognition system for languages with compound words
Patent #: 5754972
Issued on: 05/19/1998
Inventor: Baker, et al.

Method and system for natural language translation
Patent #: 5768603
Issued on: 06/16/1998
Inventor: Brown, et al.

Method and system using separate context and constituent probabilities for speech recognition in languages with compound words
Patent #: 5797122
Issued on: 08/18/1998
Inventor: Spies

Sentence processing apparatus and method thereof,utilizing dictionaries to interpolate elliptic characters or symbols Patent #: 6173253
Issued on: 01/09/2001
Inventor: Abe, et al.

Inventors

Assignee

Larnout & Hauspie Speech Products N.V.

Application

No. 295424 filed on 04/20/1999

US Classes:

704/257, Natural language704/255Specialized models

Examiners

Primary: Dorvil, Richemond

Attorney, Agent or Firm

Bromberg & Sunstein LLp

Foreign Patent References

195 10 083 DE. 09/11/1996
0 508 225 EP. 10/11/1992
0 566 848 EP. 10/11/1993
WO 93/18506 WO. 09/11/1993

International Class

G10L 015/18

Abstract

An automatic speech recognition system includes a recognition engine, a compound lexicon, and a compounder. The recognition engine generates a recognition result having a sequence of recognized words representative of an input utterance. The words in the recognition result include compound word components which may be combined to form compound words. The engine uses a recognition vocabulary of words and a language model which, for a given position in the sequence of recognized words and for selected words in the recognition vocabulary, associates a likelihood of such word occurring at such position. The compound lexicon contains a plurality of compound word components and component connecting links structured so that links are present between components that are more likely to occur together than to occur as separate words. The compounder replaces adjacent words in the sequence of recognized words which have corresponding linked components in the lexicon with a concatenation of the corresponding linked components. The probabilities in the language model minimize the likelihood of a component in the lexicon occurring in the recognition result without a corresponding linked component for concatenation.

Other References

ICASSP-95., 1995 International conference on Acoustics, Speech, and Signal Processing. Geutner, "Using morphology towars better large-vocabulary speech recognition systems". pp. 445-448, May 1995.
Spoken Language, 1996. Benton et al., "Compound words in large-vocabulary German Speech recognition systems". pp. 1165-1168. Oct. 1996.
Berton, Andre' et al, "Compound Words in Large-Vocabulary German Speech Recognition Systems", In Proc. ICSLP. 1996, Philadelphia, Pennsylvania, USA, pp. 1165-68
Spies, Marcus, "A Language Model for Compound Words in Speech Recognition", ESCA Eurospeech '95 4^th European Conference on Speech Communication and Technology, Madrid, Sep. 1995, pp. 1767-177