Determining speech recognition accuracy

• US 20040015351A1
• Filed: 07/16/2002
• Published: 01/22/2004
• Est. Priority Date: 07/16/2002
• Status: Active Grant

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Moore, Victor S., Jaiswal, Peeyush, Gandhi, Shailesh B., Toon, Gregory L.
• Granted Patent: US 7,260,534 B2
• Time in Patent Office: Days
• US Class Current: 704/240
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems

Abstract

A solution for determining the accuracy of a speech recognition system. A first graphical user interface (GUI) is provided for selecting a transaction log. The transaction log has at least one entry that specifies a speech recognition text result. A second GUI is also provided for selecting at least one audio segment corresponding to the entry. The second GUI includes an activatable icon for initiating transcription of the audio segment through a reference speech recognition engine to generate a second text result.

Claims

1. A method of determining the accuracy of a speech recognition system comprising:

providing a first graphical user interface (GUI) for selecting a transaction log wherein said transaction log has at least one entry, said entry specifying a speech recognition text result; and

providing a second GUI for selecting at least one audio segment corresponding to said entry;

wherein said second GUI comprises an activatable icon for initiating transcription of said audio segment through a reference speech recognition engine to generate a second text result.

2. The method of claim 1, wherein said second GUI comprises an input portion for receiving user corrected transcribed text.

3. The method of claim 1, further comprising:

providing a third GUI, wherein said third GUI comprises one or more controls to associate said audio segment with at least one condition.

4. The method of claim 3, wherein said condition specifies at least a person having generated said audio segment, a gender of said person, and ambient sounds influencing a recognizability of said audio segment.

5. The method of claim 3, wherein said condition is stored in said transaction log and associated with said entry.

6. The method of claim 4, wherein said ambient sounds are at least one of weather generated sound and background noise.

7. The method of claim 1 wherein, said second GUI is automatically presented upon a transaction log being selected.

8. The method of claim 1, further comprising the step of providing a fourth GUI, wherein said fourth GUI comprises one or more indicators to show an operational status of a software application used in determining the accuracy of a speech recognition system.

9. The method of claim 1, further comprising providing a fifth GUI displaying said text result and said second text result.

10. The method of claim 9, wherein said fifth GUI further displays manually entered text corresponding to said audio segment.

11. The method of claim 9, wherein said fifth GUI further displays data.

12. The method of claim 11 wherein said data is statistical data.

13. A machine-readable storage, having stored thereon a computer program having a plurality of code sections executable by a machine for causing the machine to perform the steps of:

providing a first graphical user interface (GUI) for selecting a transaction log wherein said transaction log has at least one entry, said entry specifying a speech recognition text result; and

providing a second GUI for selecting at least one audio segment corresponding to said entry;

wherein said second GUI comprises an activatable icon for initiating transcription of said audio segment through a reference speech recognition engine to generate a second text result.

14. The machine readable storage of claim 13, wherein said second GUI comprises an input portion for receiving user corrected transcribed text.

15. The machine readable storage of claim 13, further comprising:

providing a third GUI, wherein said third GUI comprises one or more controls to associate said audio segment with at least one condition.

16. The machine readable storage of claim 15, wherein said condition specifies at least a person having generated said audio segment, a gender of said person, and ambient sounds influencing a recognizability of said audio segment.

17. The machine readable storage of claim 15, wherein said condition is stored in said transaction log and associated with said entry.

18. The machine readable storage of claim 16, wherein said ambient sounds are at least one of weather generated sound and background noise.

19. The machine readable storage of claim 13, said second GUI is automatically presented upon a transaction log being selected.

20. The machine readable storage of claim 13, further comprising the step of providing a fourth GUI, wherein said fourth GUI comprises one or more indicators to show an operational status of a software application used in determining the accuracy of a speech recognition system.

21. The machine readable storage of claim 13, further comprising providing a fifth GUI displaying said text result and said second text result.

22. The machine readable storage of claim 21, wherein said fifth GUI further displays manually entered text corresponding to said audio segment.

23. The machine readable storage of claim 21, wherein said fifth GUI further displays data.

24. The machine readable storage of claim 23, wherein said data is statistical data.