Method and apparatus for speech translation with unrecognized segments

• US 5,956,668 A
• Filed: 07/18/1997
• Issued: 09/21/1999
• Est. Priority Date: 07/18/1997
• Status: Expired due to Term

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: AT&T Corporation (AT&T, Inc.)
• Inventors: Alshawi, Hiyan, Berkley, David Arthur
• Primary Examiner(s): Isen, Forester W.
• Assistant Examiner(s): EDOUARD, PATRICK NESTOR
• Application Number:
  US08/897,054
• Time in Patent Office: 795 Days
• Field of Search: 704/1, 704/2, 704/9, 704/277, 704/251, 704/253
• US Class Current: 704/2
• CPC Class Codes: G06F 40/44 Statistical methods, e.g. probability models
• CPC Class Codes: G06F 40/55 Rule-based translation
• CPC Class Codes: G10L 13/02 Methods for producing synthetic speech; Speech synthesisers
• CPC Class Codes: G10L 15/05 Word boundary detection
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

A method for the treatment of unrecognized words and phrases in spoken language translation systems is applicable to open-ended word or phrase categories, such as proper names, and involves the demarcation of one or more unrecognized segments from the input speech signal and splicing these segments into the appropriate position of the synthesized speech signal for the translation.

Claims

1. A speech translation system, comprising:

a speech recognizer, which produces based on digitized speech (i) a recognized text word sequence including one or more labels identifying segments of the digitized speech, and (ii) segment position information for said one or more labels identifying said segments of said digitized speech;

a language translation component coupled to the speech recognizer, the language translation component produces a translated text word sequence that includes said one or more labels identifying said segments of said digitized speech based on the recognized text word sequence;

a speech segment storage component coupled to the speech recognizer, the speech segment storage component stores said digitized speech and said segment position information and produces said segments of said digitized speech based on said segment position information; and

a speech synthesizer coupled to the language translation component and the speech segment storage component, the speech synthesizer produces, based on the translated text word sequence, target language output speech including said segments of said digitized speech in place of said one or more labels.

2. A system as defined in claim 1, further comprising:

a speech coder coupled to the speech recognizer and the speech segment storage component, the speech coder receives source language input speech and produces the digitized speech.

3. A method of translating speech, comprising the following steps:

(A) making from digitized speech a recognized text word sequence including one or more labels identifying segments of said digitized speech and segment position information;

(B) translating the recognized text word sequence into a translated text word sequence that includes said one or more labels identifying said segments of said digitized speech; and

(C) synthesizing target language output speech from the translated text word sequence wherein said segments of said digitized speech are substituted in place of said one or more labels based on said segment position information.

4. A method as defined in claim 3, wherein step (A) comprises the steps of:

receiving source language input speech,

making said digitized speech from the source language input speech, and

recognizing said digitized speech to make the recognized text word sequence including said one or more labels identifying said segments of said digitized speech.

5. A method as defined in claim 3, further comprising the step of:

storing segment position information for said one or more labels identifying said segments of said digitized speech.

6. A speech translation system, comprising:

means for making from digitized speech a recognized text word sequence including one or more labels identifying segments of said digitized speech and segment position information for said one or more labels;

means for translating the recognized text word sequence into a translated text word sequence that includes said one or more labels identifying said segments of said digitized speech; and

means for synthesizing target language output speech from the translated text word sequence wherein said segments of said digitized speech are substituted in place of said one or more labels based on said segment position information.