Apparatus, method, and computer program product for processing input speech

• US 8,954,333 B2
• Filed: 02/26/2008
• Issued: 02/10/2015
• Est. Priority Date: 02/27/2007
• Status: Active Grant

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation), Toshiba Digital Solutions Corporation (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Chino, Tetsuro, Kamatani, Satoshi, Furihata, Kentaro
• Primary Examiner(s): He, Jialong
• Application Number:
  US12/037,724
• Publication Number:
  US 20080208597A1
• Time in Patent Office: 2,541 Days
• Field of Search: 704 2- 5, 704/7
• US Class Current: 704/275
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding

Abstract

An analyzing unit performs a morphological analysis of an input character string that is obtained by processing input speech. A generating unit divides the input character string in units of division previously decided, that is composed of one or plural morphemes, and generates partial character strings including part of components of the divided input character string. A candidate output unit outputs the generated partial character strings to a display unit. A selection receiving unit receives a partial character string selected from the outputted partial character strings as a target to be processed.

Claims

1. A speech processing apparatus comprising:

a speech receiving unit that receives input speech;

a speech processing unit that performs a speech recognizing process on the input speech to obtain a recognition candidate;

an analyzing unit that performs a morphological analysis of the recognition candidate;

a generating unit that;

divides, via a processor of the apparatus, the recognition candidate into a plurality of first segments each including at least one morpheme, and

generates a plurality of partial character string candidates, each consisting of a subset of the first segments, the subset being obtained by removing one or more of the first segments from the recognition candidate, the subset for each partial character string candidate being different from the other subsets;

a first output unit that outputs the partial character string candidates to a display unit; and

a selection receiving unit that receives a selection of a partial character string selected from the partial character string candidates.

2. The apparatus according to claim 1, wherein the speech receiving unit receives the input speech in a first language, and wherein the apparatus further comprises:

a translating unit that translates the received partial character string selection into a second language to obtain a translation result; and

a second output unit that outputs the translation result.

4. The apparatus according to claim 1, wherein the generating unit:

divides the recognition candidate into a plurality of second segments representing a syntactic structure unit of a sentence including a word, a segment and a phrase, and

generates the partial character string candidates so as to include at least one of the second segments.

5. The apparatus according to claim 1, wherein the generating unit:

divides the recognition candidate into a plurality of third segments representing a semantic unit of a phrase including at least one of a figure, a time, a degree, a greeting and a formulaic expression and

generates the partial character string candidates obtained by removing at least one of the third segments from the recognition candidate.

6. The apparatus according to claim 1, further comprising:

a storage unit that can store the received partial character string and the recognition candidate as a generation source of the partial character string, in association with each other,

wherein the selection receiving unit stores in the storage unit the received partial character string and the recognition candidate as the generation source of the partial character string, in association with each other.

7. The apparatus according to claim 6, further comprising:

a determining unit that determines whether the partial character string corresponding to the recognition candidate is stored in the storage unit,

wherein the generating unit obtains the partial character string corresponding to the recognition candidate from the storage unit to generate the partial character string candidate, when the partial character string corresponding to the recognition candidate is stored in the storage unit.

8. The apparatus according to claim 6, further comprising:

a determining unit that determines whether the partial character string corresponding to the recognition candidate is stored in the storage unit,

wherein the first output unit prioritizes the output of the partial character string candidate stored in the storage unit over the partial character string candidates that are not stored in the storage unit.

9. The apparatus according to claim 1, wherein the speech processing unit further calculates likelihood indicating a probability of the recognition candidate of the received input speech, and the apparatus further comprises a determining unit that determines whether the likelihood is smaller than a predetermined threshold value, wherein the generating unit generates the partial character string candidates when the determining unit determines that the likelihood is smaller than the threshold value.

10. The apparatus according to claim 9, wherein the first output unit outputs the recognition candidate when the determining unit determines that the likelihood is larger than the threshold value.

11. The apparatus according to claim 1, wherein the first output unit extracts a predetermined number of the partial character string candidates from the partial character string candidates, and outputs the extracted partial character string candidates.

12. The apparatus according to claim 1, wherein:

the speech processing unit further calculates likelihood indicating a probability of the recognition candidate of the received input speech, and

the first output unit outputs a predetermined number of partial character string candidates among the partial character string candidates according to the likelihood of the recognition candidate.

3. A speech processing apparatus comprising:

a speech receiving unit receives the input speech in a first language;

a speech processing unit performs a speech recognizing process on the received input speech to obtain a recognition candidate, and translates the recognition candidate into a second language to obtain a text character string;

an analyzing unit that performs a morphological analysis of the text character string;

a generating unit that;

divides, via a processor of the apparatus, the text character string into a plurality of segments each including at least one morpheme, and

generates a plurality of partial character string candidates, each consisting of a subset of the segments, the subset being obtained by removing one or more of the segments from the text character string, the subset for each partial character string candidate being different from the other subsets;

a first output unit that outputs the partial character string candidates to a display unit; and

a selection receiving unit that receives a selection of a partial character string selected from the partial character string candidates.

13. A speech processing method comprising:

receiving an input speech;

performing a speech recognition process on the input speech to obtain a recognition candidate;

performing a morphological analysis of the recognition candidate;

dividing the recognition candidate into a plurality of segments each including at least one morpheme;

generating, via execution of a processor, a plurality of partial character string candidates, each consisting of a subset of the segments, the subset being obtained by removing one or more of the segments from the recognition candidate, the subset for each partial character string candidate being different from the other subsets;

outputting the partial character string candidates to a display unit; and

receiving a selection of a partial character string selected from the partial character string candidates.

14. A tangible non-transitory computer-readable medium including programmed instructions for processing an input speech, wherein the instructions, when executed by a computer, cause the computer to perform:

receiving an input speech;

performing a speech recognizing process on the input speech to obtain a recognition candidate;

performing a morphological analysis of the recognition candidate;

dividing the recognition candidate into a plurality of segments each including at least one morpheme;

generating, via execution of a processor, a plurality of partial character string candidates, each consisting of a subset of the segments, the subset being obtained by removing one or more of the segments from the recognition candidate, the subset for each partial character string candidate being different from the other subsets;

outputting the partial character string candidates to a display unit; and

receiving a partial character string selected from the partial character string candidates.