System and method for generating videoconference transcriptions

• US 8,630,854 B2
• Filed: 08/31/2010
• Issued: 01/14/2014
• Est. Priority Date: 08/31/2010
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Marvit, David L.
• Primary Examiner(s): Neway, Samuel G
• Application Number:
  US12/872,268
• Publication Number:
  US 20120053936A1
• Time in Patent Office: 1,232 Days
• Field of Search: 348 1408- 1409, 704246-250
• US Class Current: 704/246
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 17/08 Use of distortion metrics or a particular distance between probe pattern and reference templates
• CPC Class Codes: G10L 17/10 Multimodal systems, i.e. based on the integration of multiple recognition engines or fusion of expert systems
• CPC Class Codes: H04N 7/147 Communication arrangements, e.g. identifying the communication as a video-communication, intermediate storage of the signals selecting H04Q

Abstract

A method for generating a transcription of a videoconference includes matching human speech of a videoconference to writable symbols. The human speech is encoded in audio data of the videoconference. The writable symbols are parsed into a plurality of statements. For each statement of the plurality of statements, user profile data stored in computer-readable memory is used to determine which participant of a plurality of participants of the videoconference is most likely the source of the statement. A transcription of the videoconference is generated that identifies for each statement the determination of which participant of the plurality of participants of the videoconference is most likely the source of the statement.

Claims

1. A method for generating a transcription of a videoconference, comprising:

matching human speech of a videoconference to writable symbols, the human speech encoded in audio data of the videoconference;

determining a probability that a portion of the human speech matches a profile of a participant of a plurality of participants of the videoconference, the profile stored in tangible computer-readable memory;

if the probability is less than a predetermined threshold, using video data of the videoconference to determine which participant of the plurality of participants of the videoconference is the most likely source of the portion of the human speech; and

generating a transcription of the videoconference that identifies an association of the portion of the human speech and the participant of the plurality of participants of the videoconference determined to be the most likely source of the portion of human speech.

2. The method of claim 1, wherein the profile comprises voice profile data of the participant of the plurality of participants of the videoconference.

3. The method of claim 1, wherein the profile comprises visual profile data of the participant of the plurality of participants of the videoconference.

4. The method of claim 1, further comprising generating at least a portion of the profile during the videoconference.

5. The method of claim 1, further comprising generating at least a portion of the profile before the videoconference begins.

6. The method of claim 1, further comprising generating at least a portion of the profile after the videoconference has terminated.

15. The method of claim 1, wherein the profile comprises a location of the participant.

16. The method of claim 1, wherein the profile comprises an address of the participant.

7. A non-transitory computer-readable memory storing logic, the logic operable when executed by one or more processors to:

match human speech of a videoconference to writable symbols, the human speech encoded in audio data of the videoconference;

determine a probability that a portion of the human speech matches a profile of a participant of a plurality of participants of the videoconference, the profile stored in tangible computer-readable memory;

if the probability is less than a predetermined threshold, use video data of the videoconference to determine which participant of the plurality of participants of the videoconference is the most likely source of the portion of the human speech; and

generate a transcription of the videoconference that identifies for each statement the determination of which participant of the plurality of participants of the videoconference is most likely the source of the statement.

8. The non-transitory computer-readable memory of claim 7, wherein the profile comprises voice profile data of the participant of the plurality of participants of the videoconference.

9. The non-transitory computer-readable memory of claim 7, wherein the profile comprises visual profile data of the participant of the plurality of participants of the videoconference.

10. The non-transitory computer-readable memory of claim 7, wherein the logic is further operable when executed by the one or more processors to generate at least a portion of the profile during the videoconference.

11. The non-transitory computer-readable memory of claim 7, wherein the logic is further operable when executed by the one or more processors to generate at least a portion of the profile before the videoconference begins.

12. The non-transitory computer-readable memory of claim 7, wherein the logic is further operable when executed by the one or more processors to generate at least a portion of the profile after the videoconference has terminated.

17. The non-transitory computer readable memory of claim 7, wherein the profile comprises a location of the participant.

18. The non-transitory computer readable memory of claim 7, wherein the profile comprises an address of the participant.

13. A method for generating a transcription of a videoconference, comprising:

matching human speech of a videoconference to writable symbols, the human speech encoded in an audio data stream of the videoconference;

determining a probability that a portion of the human speech matches a voice profile of a participant of a plurality of participants of the videoconference, the voice profile stored in tangible computer-readable memory;

if the probability is less than a predetermined threshold, using video data of the videoconference to determine which participant of the plurality of participants of the videoconference is the most likely source of the portion of the human speech, the video data corresponding to the portion of the human speech; and

generating a transcription of the videoconference that identifies an association of the portion of the human speech and the participant of the plurality of participants determined to be the most likely source of the portion of the human speech.

14. The method of claim 13, further comprising generating the voice profile using the audio data stream of the videoconference.

19. The method of claim 13, wherein the voice profile comprises a location of the participant.

20. The method of claim 13, wherein the voice profile comprises an address of the participant.