Systems and methods for hands-free voice control and voice search

• US 8,700,399 B2
• Filed: 07/06/2010
• Issued: 04/15/2014
• Est. Priority Date: 07/06/2009
• Status: Active Grant

• Current Assignee: Sensory Incorporated
• Original Assignee: Sensory Incorporated
• Inventors: Vermeulen, Pieter J., Shaw, Jonathan, Mozer, Todd F.
• Primary Examiner(s): WOZNIAK, JAMES S
• Application Number:
  US12/831,051
• Publication Number:
  US 20110166855A1
• Time in Patent Office: 1,379 Days
• Field of Search: 704/231, 704/236, 704/242, 704/251, 704/254, 704/248
• US Class Current: 704/242
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: G10L 2015/081 Search algorithms, e.g. Baum-Welch or Viterbi
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

In one embodiment the present invention includes a method comprising receiving an acoustic input signal and processing the acoustic input signal with a plurality of acoustic recognition processes configured to recognize the same target sound. Different acoustic recognition processes start processing different segments of the acoustic input signal at different time points in the acoustic input signal. In one embodiment, initial states in the recognition processes may be configured on each time step.

Claims

1. A method comprising:

receiving, in a processor, an acoustic input signal; and

processing the acoustic input signal with a plurality of acoustic recognition processes to recognize a predetermined target sound within the acoustic input signal, the acoustic input signal being temporally divided into a plurality of frames, the plurality of acoustic recognition processes being implemented using a single Viterbi search of a plurality of states corresponding to acoustic units of an acoustic model of the predetermined target sound, the plurality of states including initial states and final states of the acoustic model, wherein

the initial states are reset on each frame if a score for the initial state on a previous frame is below a first threshold,

said score is calculated for a plurality of said states on each frame, the calculating comprising increasing the score by a per frame offset so that scores of different durations are comparable, and

a result is generated when a score of a final state increases above a second threshold.

2. The method of claim 1 wherein the score for an initial state is reset to a predetermined value on each frame if the score is less than the first threshold before calculating a current score for the initial state based on a received acoustic unit for a current frame.

3. The method of claim 2 wherein said predetermined value is a constant.

4. The method of claim 2 wherein different initial states are reset to different predetermined values.

5. The method of claim 1 wherein the offset is a constant.

6. The method of claim 1 wherein different states have different associated offsets.

7. The method of claim 1 wherein the plurality of acoustic recognition processes operate on multiple words in parallel.

8. A non-transitory computer readable storage medium having stored thereon program code executable by a processor, said program code comprising:

code that causes the processor to receive an acoustic input signal; and

code that causes the processor to process the acoustic input signal with a plurality of acoustic recognition processes to recognize a predetermined target sound within the acoustic input signal, the acoustic input signal being temporally divided into a plurality of frames, the plurality of acoustic recognition processes being implemented using a single Viterbi search of a plurality of states corresponding to acoustic units of an acoustic model of the predetermined target sound, the plurality of states including initial states and final states of the acoustic model, wherein

the initial states are reset on each frame if a score for the initial state on a previous frame is below a first threshold,

said score is calculated for a plurality of said states on each frame, the calculating comprising increasing the score by a per frame offset so that scores of different durations are comparable, and

a result is generated when a score of a final state increases above a second threshold.

9. The non-transitory computer readable medium of claim 8 wherein the score for an initial state is reset to a predetermined value on each frame if the score is less than the first threshold before calculating a current score for the initial state based on a received acoustic unit for a current frame.

10. The non-transitory computer readable medium of claim 9 wherein the predetermined value is a constant.

11. The non-transitory computer readable medium of claim 9 wherein different initial states are reset to different predetermined values.

12. The non-transitory computer readable medium of claim 8 wherein the offset is a constant.

13. The non-transitory computer readable medium of claim 8 wherein different states have different associated offsets.

14. The non-transitory computer readable medium of claim 8 wherein the plurality of acoustic recognition processes operate on multiple words in parallel.

15. A computer system comprising:

a processor; and

a non-transitory computer readable medium having stored thereon instructions that, when executed by the processor, causes the processor to;

receive an acoustic input signal; and

process the acoustic input signal with a plurality of acoustic recognition processes to recognize a predetermined target sound within the acoustic input signal, the acoustic input signal being temporally divided into a plurality of frames, the plurality of acoustic recognition processes being implemented using a single Viterbi search of a plurality of states corresponding to acoustic units of an acoustic model of the predetermined target sound, the plurality of states including initial states and final states of the acoustic model, wherein

the initial states are reset on each frame if a score for the initial state on a previous frame is below a first threshold,

said score is calculated for a plurality of said states on each frame, the calculating comprising increasing the score by a per frame offset so that scores of different durations are comparable, and

a result is generated when a score of a final state increases above a second threshold.

16. The computer system of claim 15 wherein the score for an initial state is reset to a predetermined value on each frame if the score is less than the first threshold before calculating a current score for the initial state based on a received acoustic unit for a current frame.

17. The computer system of claim 16 wherein the predetermined value is a constant.

18. The computer system of claim 16 wherein different initial states are reset to different predetermined values.

19. The computer system of claim 15 wherein the offset is a constant.

20. The computer system of claim 15 wherein different states have different associated offsets.

21. The computer system of claim 15 wherein the plurality of acoustic recognition processes operate on multiple words in parallel.