Speech recognition method and apparatus with automatic parameter selection based on hardware running environment

• US 5,717,820 A
• Filed: 03/07/1995
• Issued: 02/10/1998
• Est. Priority Date: 03/10/1994
• Status: Expired due to Term

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Kimura, Shinta, Hamasaki, Ryosuke
• Primary Examiner(s): MacDonald, Allen R.
• Assistant Examiner(s): MATTSON, ROBERT
• Application Number:
  US08/399,631
• Time in Patent Office: 1,071 Days
• Field of Search: 395/2.4, 395/2.84, 395/2.6
• US Class Current: 704/231
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 15/28 Constructional details of speech recognition systems
• CPC Class Codes: G10L 2015/0631 Creating reference templates; Clustering
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A speech recognition method and apparatus enable the use of the same recognition software for speaker dependent recognition under various running environments such as different computers. The speech recognition method operates in the form of computer software to detect at least one running environment of hardware. The hardware running environment may include processing speed of a computer. Recognition parameters are then determined which correspond to the detected running environment. Next, the speech recognition method converts an input analog speech signal into a digital speech signal. The digital speech signal, in the form of extracted feature data, is then compared with digital feature data in a dictionary. The speech is then recognized in response to a match of the data.

Claims

1. A speech recognition method which carries out a speech recognition process on an input speech signal based on recognition parameters using recognition software, said speech recognition method comprising the steps of:

(a) detecting at least one running environment provided by hardware which is to run the recognition software;

(b) determining recognition parameters depending on the running environment detected by said step (a); and

(c) carrying out the speech recognition process on the input speech signal based on the recognition parameters determined by said step (b) using the recognition software,

said recognition parameters determining a quality of the speech recognition process.

2. The speech recognition method as claimed in claim 1, which further comprises the steps of:

(d) extracting feature data of the input speech signal depending on the recognition parameters determined by said step (b),

said step (c) carrying out the speech recognition by matching the feature data extracted by said step (d) with feature data registered in one or a plurality of dictionaries.

3. The speech recognition method as claimed in claim 2, which further comprises the steps of:

(e) registering the feature data of the speech signal extracted by said step (d) in the one or plurality of dictionaries depending on the recognition parameters determined by said step (b).

4. The speech recognition method as claimed in claim 2, which further comprises the steps of:

(f) specifying one dictionary out of a plurality of dictionaries, based on the recognition parameters,

said step (c) carrying out the speech recognition by matching the feature data extracted by said step (d) with feature data registered in the dictionary specified by said step (f).

5. The speech recognition method as claimed in claim 2, which further comprises the steps of:

(g) measuring a running time or processing time of a performance measuring program which measures a performance of the hardware by starting the performance measuring program,

said step (b) determining the recognition parameters by taking into consideration the running time or processing time measured by said step (g).

6. The speech recognition method as claimed in claim 2, which further comprises the steps of:

(g) measuring a running time or processing time of a performance measuring program which measures a performance of the hardware by starting the performance measuring program,

said step (d) extracting the feature data of the input speech signal using an analyzing technique that is changed depending on the running time or processing time measured by said step (g).

7. The speech recognition method as claimed in claim 2, which further comprises the steps of:

(h) measuring a time required to extract the feature data of the input speech signal by said step (d),

said step (b) determining the recognition parameters by taking into consideration the time measured by said step (h).

8. The speech recognition method as claimed in claim 2, which further comprises the steps of:

(h) measuring a time required to extract the feature data of the input speech signal by said step (d),

said step (d) extracting the feature data of the input speech signal using an analyzing technique that is changed depending on the time measured by said step (h).

9. The speech recognition method as claimed in claim 1, wherein said step (b) determines the recognition parameters depending on the running environment detected by said step (a) by referring to a conversion table which manages corresponding relationships of running environment values and recognition parameter values.

10. The speech recognition method as claimed in claim 9, wherein said step (b) determines the recognition parameters to be used by said step (c) upon the speech recognition process by carrying out a predetermined operation on a predetermined combination of the recognition parameter values which are specified by the detected running environment values and are read from the conversion table.

11. A speech recognition apparatus which carries out a speech recognition process on an input speech signal based on recognition parameters using recognition software, said speech recognition apparatus comprising:

detecting means for detecting at least one running environment provided by hardware which is to run the recognition software;

determining means, coupled to said detecting means, for determining recognition parameters depending on the running environment detected by said detecting means; and

matching means, coupled to said determining means, for carrying out the speech recognition process on the input speech signal based on the recognition parameters determined by said determining means using the recognition software,

said recognition parameters determining a quality of the speech recognition process.

12. The speech recognition apparatus as claimed in claim 11, which further comprises:

one or a plurality of dictionaries, coupled to said matching means, having feature data registered therein; and

analyzing means, coupled to said determining means, for extracting feature data of the input speech signal depending on the recognition parameters determined by said determining means,

said matching means carrying out the speech recognition by matching the feature data extracted by said analyzing means with feature data registered in said one or plurality of dictionaries.

13. The speech recognition apparatus as claimed in claim 12, which further comprises:

registering means, coupled to said one or plurality of dictionaries, for registering the feature data of the speech signal extracted by said analyzing means in said one or plurality of dictionaries depending on the recognition parameters determined by said determining means.

14. The speech recognition apparatus as claimed in claim 12, which further comprises:

specifying means, coupled to a plurality of dictionaries, for specifying one dictionary out of the plurality of dictionaries based on the recognition parameters,

said matching means carrying out the speech recognition by matching the feature data extracted by said analyzing means with feature data registered in the dictionary specified by said specifying means.

15. The speech recognition apparatus as claimed in claim 12, which further comprises:

measuring means for measuring a running time or processing time of a performance measuring program which measures performance of the hardware by starting the performance measuring program,

said determining means determining the recognition parameters by considering the running time or processing time measured by said measuring means.

16. The speech recognition apparatus as claimed in claim 12, which further comprises:

measuring means for measuring a running time or processing time of a performance measuring program which measures a performance of the hardware by starting the performance measuring program,

said analyzing means extracting the feature data of the input speech signal using an analyzing technique that is changed depending on the running time or processing time measured by said measuring means.

17. The speech recognition apparatus as claimed in claim 12, which further comprises:

measuring means for measuring a time required to extract the feature data of the input speech signal by said analyzing means,

said determining means determining the recognition parameters by considering the time measured by said measuring means.

18. The speech recognition apparatus as claimed in claim 12, which further comprises:

measuring means for measuring a time required to extract the feature data of the input speech signal by said analyzing means,

said analyzing means extracting the feature data of the input speech signal using an analyzing technique that is changed depending on the time measured by said measuring means.

19. The speech recognition apparatus as claimed in claim 11, which further comprises:

a conversion table managing corresponding relationships of each of running environment values and each of recognition parameter values,

said determining means determining the recognition parameters depending on the running environment detected by said detecting means by referring to said conversion table.

20. The speech recognition apparatus as claimed in claim 19, wherein said determining means determines the recognition parameters to be used by said matching means upon the speech recognition process by carrying out a predetermined operation on a predetermined combination of the recognition parameter values which are specified by the detected running environment values and are read from said conversion table.

21. A storage unit adapted for use on a speech recognition apparatus which carries out a speech recognition process on an input speech signal based on recognition parameters using recognition software, said speech recognition apparatus comprising:

detecting means for detecting at least one running environment provided by hardware which is to run the recognition software;

determining means, coupled to said detecting means, for determining recognition parameters depending on the running environment detected by said detecting means; and

matching means, coupled to said determining means, for carrying out the speech recognition process on the input speech signal based on the recognition parameters determined by said determining means using the recognition software,

said storage unit comprising;

values of running environments defined as running environment values;

values of recognition parameters defined as recognition parameter values; and

a conversion table managing corresponding relationships of each of the running environment values and each of the recognition parameter values, so that said determining means determines the recognition parameters depending on the running environment detected by said detecting means by referring to said conversion table,

said recognition parameters determining a quality of the speech recognition process.

22. The storage unit as claimed in claim 21, wherein the recognition software is stored together with said conversion table.

23. The storage unit as claimed in claim 22, which further comprises:

one or a plurality of dictionaries having registered feature data stored with said conversion table.

24. The storage unit as claimed in claim 22, which further comprises:

a plurality of dictionaries having registered feature data stored with said conversion table, said plurality of dictionaries respectively having recognition parameter dependent registered feature data.