SPEECH RECOGNITION SYSTEM AND METHOD BASED ON WORD-LEVEL CANDIDATE GENERATION

• US 20120290303A1
• Filed: 05/08/2012
• Published: 11/15/2012
• Est. Priority Date: 05/12/2011
• Status: Active Grant

• Current Assignee: NHN Corporation
• Original Assignee: NHN Corporation
• Inventors: LEE, Sang Ho, KIM, Hoon, KOO, Dong Ook, JUNG, Dae Sung
• Granted Patent: US 9,002,708 B2
• Time in Patent Office: Days
• US Class Current: 704/251
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

A speech recognition system and method based on word-level candidate generation are provided. The speech recognition system may include a speech recognition result verifying unit to verify a word sequence and a candidate word for at least one word included in the word sequence when the word sequence and the candidate word are provided as a result of speech recognition. A word sequence displaying unit may display the word sequence in which the at least one word is visually distinguishable from other words of the word sequence. The word sequence displaying unit may display the word sequence by replacing the at least one word with the candidate word when the at least one word is selected by a user.

Claims

1. A speech recognition system, comprising:

a speech recognition result verifying unit configured to verify a speech recognition result comprising a word sequence and at least one candidate word for at least one word in the word sequence; and

a word sequence displaying unit configured to display the word sequence, the at least one word having the at least one candidate word being visually distinguishable from other words of the word sequence,

wherein, in response to the at least one word being selected, the word sequence displaying unit is configured to display the word sequence by replacing the at least one word with the at least one candidate word.

2. The system of claim 1, further comprising:

a user interface configured to delete the word sequence in response to receiving a command through the user interface.

3. The system of claim 1, wherein a signature indicating that the word sequence is provided through a speech recognition process is displayed when the word sequence is displayed through the display device, and

wherein the signature indicating that the word sequence is provided through the speech recognition process is transmitted along with the word sequence when the word sequence is transmitted to another device.

4. The system of claim 1, further comprising:

a sound signal transmitting unit configured to transmit, to a speech recognition server, a sound signal input through an input device; and

a speech recognition result receiving unit configured to receive, from the speech recognition server, the speech recognition result corresponding to the sound signal.

5. The system of claim 4, wherein the speech recognition server comprises:

a speech recognition result generating unit configured to generate the speech recognition result corresponding to a sound signal input through an input device.

6. The system of claim 1, wherein the word sequence is selected as a word sequence having a highest probability among probabilities of a plurality of word sequences matching a received sound signal.

7. The system of claim 1, wherein the at least one candidate word is selected among at least one recognized word,

wherein the at least one candidate word is recognized during a time period determined based on a time when recognition of the at least one recognized word is initiated and a time when recognition of the at least one recognized word is terminated.

8. The system of claim 7, wherein the at least one candidate word is selected based on confidences that are calculated for each of the at least one recognized word.

9. The system of claim 7, wherein the time period for selecting the at least one candidate word is determined based on the time when the recognition of the at least one recognized word is initiated, the time when the recognition of the at least one recognized word is terminated, and a predetermined allowable time value added to the time when the recognition of the at least one recognized word is terminated or subtracted from the time when the recognition of the at least one recognized word is initiated.

10. A speech recognition system, comprising:

a speech recognition result generating unit configured to determine a word sequence corresponding to a sound signal and at least one candidate word for at least one word in the word sequence, and to generate a speech recognition result comprising the word sequence and the at least one candidate word; and

a speech recognition result providing unit configured to provide the speech recognition result to a terminal,

wherein

the word sequence is displayed on the terminal,

the at least one word is visually distinguishable from other words of the word sequence, and

in response to selection of the at least one word, the word sequence is displayed by replacing the at least one word with the at least one candidate word.

11. The system of claim 10, wherein the terminal comprises a user interface configured to delete the word sequence in response to receiving a command via the user interface.

12. The system of claim 10, wherein a signature indicating that the word sequence is provided through a speech recognition process is displayed when the word sequence is displayed in the terminal, and

wherein the signature indicating that the word sequence is provided through the speech recognition process is transmitted along with the word sequence when the word sequence is transmitted to another device.

13. The system of claim 10, further comprising:

a sound signal receiving unit configured to receive, from the terminal, the sound signal input through an input device of the terminal.

14. The system of claim 10, wherein the speech recognition result generating unit selects, as the word sequence, a word sequence having a highest probability among probabilities of a plurality of word sequences matching the sound signal.

15. The system of claim 10, wherein the speech recognition result generating unit selects the at least one candidate word among at least one recognized word,

wherein the at least one candidate word is recognized during a time period determined based on a time when recognition of the at least one recognized word is initiated and a time when recognition of the at least one recognized word is terminated.

16. The system of claim 15, wherein the at least one candidate word is selected based on confidences that are calculated for each of the at least one recognized word.

17. The system of claim 15, wherein the time period for selecting the at least one candidate word is determined based on the time when the recognition of the at least one recognized word is initiated, the time when the recognition of the at least one recognized word is terminated, and a predetermined allowable time value added to the time when the recognition of the at least one recognized word is terminated or subtracted from the time when the recognition of the at least one recognized word is initiated.

18. A speech recognition system, comprising:

a speech recognition result verifying unit configured to verify a result comprising a word sequence and at least one candidate word for at least one word in the word sequence; and

a word sequence displaying unit configured to display the word sequence in which the at least one word is visually distinguishable from other words of the word sequence,

wherein the word sequence displaying unit displays a list including the at least one candidate word in response to the at least one word being selected.

19. A speech recognition system, comprising:

a speech recognition result generating unit configured to determine a word sequence corresponding to a sound signal and at least one candidate word for at least one word in the word sequence, and to generate a speech recognition result comprising the word sequence and the candidate word; and

a speech recognition result providing unit configured to provide the speech recognition result to a terminal,

wherein

the word sequence is displayed on the terminal,

the at least one word is visually distinguishable from other words of the word sequence, and

in response to the at least one word being selected, the terminal displays a list comprising the at least one candidate word.

20. A speech recognition method, comprising:

verifying, by a processor, a word sequence and at least one candidate word for at least one word in the word sequence in response to receiving a speech recognition result comprising the word sequence and the at least one candidate word; and

displaying, by a display unit connected to the processor, the word sequence in which the at least one word is visually distinguishable from other words of the word sequence,

wherein the displaying comprises displaying the word sequence by replacing the at least one word with the at least one candidate word in response to the at least one word being selected.

24. A non-transitory computer-readable storage medium comprising a program, that when executed, instructs a processor to perform the method of claim 20.

21. A speech recognition method, comprising:

determining, by a processor, a word sequence corresponding to a sound signal and at least one candidate word for at least one word in the word sequence;

generating, by the processor, the word sequence and the at least one candidate word as a speech recognition result; and

providing the speech recognition result to a terminal,

wherein

the word sequence is displayed on the terminal,

the at least one word being visually distinguishable from other words of the word sequence, and

in response to selection of the at least one word, the word sequence is displayed by replacing the at least one word with the at least one candidate word.

22. A speech recognition method, comprising:

verifying, by a processor, a word sequence and at least one candidate word for at least one word in the word sequence in response to receiving a speech recognition result comprising the word sequence and the candidate word; and

displaying, by a display unit connected to the processor, the word sequence in which the at least one word is visually distinguishable from other words of the word sequence,

wherein the displaying comprises displaying a list comprising the at least one candidate word in response to selection of the at least one word.

23. A speech recognition method, comprising:

determining, by a processor, a word sequence corresponding to a sound signal and a candidate word for at least one word in the word sequence;

generating, by the processor, a speech recognition result comprising the word sequence and the candidate word; and

providing the speech recognition result to a terminal,

wherein

the word sequence is displayed on the terminal,

the at least one word is visually distinguishable from other words of the word sequence, and

a list comprising the at least one candidate word is displayed in response to a selection of the at least one word.