Method and system for extracting and utilizing metadata to improve accuracy in speech to text conversions

• US 20060074895A1
• Filed: 09/29/2004
• Published: 04/06/2006
• Est. Priority Date: 09/29/2004
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Belknap, William Russell
• Granted Patent: US 7,908,141 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/60 of audio data
• CPC Class Codes: G06F 16/685 using automatically derived transcript of audio data, e.g. lyrics speech recognition G10L15/00

Abstract

A computer-based system and method for speech to text conversion preprocessing of a presentation with a speech audio, useable in real time. The method captures a presentation speech audio input to be converted into text, temporally associates the speech audio input with at least one supporting text source from the same presentation containing common keywords and creates an optimized and prioritized keyword positional index metadata set for inputting into a speech to text conversion processor.

Claims

1. A computer-based method for speech to text conversion preprocessing of a presentation with a speech audio, comprising:

(a) capturing a presentation speech audio input to be converted into text;

(b) temporally associating the speech audio input with at least one supporting text source from the same presentation containing common keywords; and

(c) creating an optimized and prioritized keyword positional index metadata set for inputting into a speech to text conversion processor.

2. The method according to claim 1, wherein the keyword positional index metadata set comprises a temporally-based keyword vocabulary set of an optimized keyword dictionary.

3. The method according to claim 1, wherein the step of creating the keyword positional index metadata set further comprises:

a) extracting words from the supporting text source, b) limiting words to keywords, c) prioritizing the keywords, and d) temporally associating each keyword temporal location with its position in the speech audio input, for creating the keyword positional index metadata set.

4. The method according to claim 3, further comprising the step of further increasing the keyword priority within the speech audio input segment when the keyword temporal location is correlated with the keyword audio position.

5. The method according to claim 1, wherein the method is used in real time.

6. The method according to claim 1, wherein the presentation is selected from the group comprising audio-visual multimedia presentation, audio, video, text and graphics presentation, live speaker presentation, telephone conference, a recording, slides, video close caption, images and computer files.

7. The method according to claim 1, wherein the supporting text source is selected from the group comprising audio, video, text, graphics, slides, video close caption, images and computer files.

8. A computer-based system for speech to text conversion preprocessing of a presentation with a speech audio, comprising means for performing the steps of:

(a) capturing a presentation speech audio input to be converted into text;

(b) temporally associating the speech audio input with at least one supporting text source from the same presentation containing common keywords; and

(c) creating an optimized and prioritized keyword positional index metadata set for inputting into a speech to text conversion processor.

9. The system according to claim 8, wherein the keyword positional index metadata set comprises a temporally-based keyword vocabulary set of an optimized keyword dictionary.

10. The system according to claim 8, wherein the step of creating the keyword positional index metadata set further comprises means for:

a) extracting words from the supporting text source, b) limiting words to keywords, c) prioritizing the keywords, and d) temporally associating each keyword temporal location with its position in the speech audio input, for creating the keyword positional index metadata set.

11. The system according to claim 10, further comprising the means for further increasing the keyword priority within the speech audio input segment when the keyword temporal location is correlated with the keyword audio position.

12. The system according to claim 8, wherein the system is used in real time.

13. The system according to claim 8, wherein the presentation is selected from the group comprising audio-visual multimedia presentation, audio, video, text and graphics presentation, live speaker presentation, telephone conference, a recording, slides, video close caption, images and computer files.

14. The system according to claim 8, wherein the supporting text source is selected from the group comprising audio, video, text, graphics, slides, video close caption, images and computer files.

15. A computer usable medium, in a database management system, tangibly embodying a program of instructions executable by the computer to perform a computer-based method for speech to text conversion preprocessing of a presentation with a speech audio, comprising:

(a) capturing a presentation speech audio input to be converted into text;

(b) temporally associating the speech audio input with at least one supporting text source from the same presentation containing common keywords; and

(c) creating an optimized and prioritized keyword positional index metadata set for inputting into a speech to text conversion processor.

16. The method according to claim 15, wherein the keyword positional index metadata set comprises a temporally-based keyword vocabulary set of an optimized keyword dictionary.

17. The method according to claim 15, wherein the step of creating the keyword positional index metadata set further comprises:

a) extracting words from the supporting text source, b) limiting words to keywords, c) prioritizing the keywords, and d) temporally associating each keyword temporal location with its position in the speech audio input, for creating the keyword positional index metadata set.

18. The method according to claim 17, further comprising the step of further increasing the keyword priority within the speech audio input segment when the keyword temporal location is correlated with the keyword audio position.

19. The method according to claim 15, wherein the method is used in real time.

20. The method according to claim 15, wherein the presentation is selected from the group comprising audio-visual multimedia presentation, audio, video, text and graphics presentation, live speaker presentation, telephone conference, a recording, slides, video close caption, images and computer files.

21. The method according to claim 15, wherein the supporting text source is selected from the group comprising audio, video, text, graphics, slides, video close caption, images and computer files.