On-line voice translation method and device

• US 9,910,851 B2
• Filed: 11/12/2014
• Issued: 03/06/2018
• Est. Priority Date: 12/25/2013
• Status: Active Grant

• Current Assignee: Beijing Baidu Netcom Science and Technology Company Limited (Baidu Incorporated)
• Original Assignee: Beijing Baidu Netcom Science and Technology Company Limited (Baidu Incorporated)
• Inventors: Wang, Haifeng, Wu, Hua
• Primary Examiner(s): BAKER, MATTHEW H
• Application Number:
  US14/893,008
• Publication Number:
  US 20160147744A1
• Time in Patent Office: 1,210 Days
• CPC Class Codes: G06F 40/49 using very large corpora, e.g. the web
• CPC Class Codes: G06F 40/51 Translation evaluation
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
• CPC Class Codes: G10L 15/005 Language recognition
• CPC Class Codes: G10L 15/01 Assessment or evaluation of speech recognition systems

Abstract

Disclosed are on-line voice translation method and device. The method comprises: conducting voice recognition on first voice information input by a first user, so as to obtain first recognition information; prompting the first user to confirm the first recognition information; translating the confirmed first recognition information to obtain and output first translation information; extracting, according to second information which is fed back by a second user, associated information corresponding to the second information; and correcting the first translation information according to the associated information and outputting the corrected translation information. By means of the on-line voice translation method and device, smooth communication can be ensured in cross-language exchanges.

Claims

1. An on-line voice translation method, comprising:

conducting voice recognition on first voice information input by a first user, so as to obtain first recognition information;

determining a degree of confidence of the first recognition information;

prompting the first user to confirm the first recognition information according to the degree of confidence of the first recognition information;

translating the confirmed first recognition information to obtain and output first translation information;

extracting, according to second information which is fed back by a second user, rectification information corresponding to the second information; and

correcting the first translation information according to the rectification information and outputting the corrected translation information,

wherein determining the degree of confidence of the first recognition information comprises;

determining the degree of confidence of the first recognition information according to a grammar rule in a current dialogue scene, wherein the degree of confidence of the first recognition information is determined to be high when the first recognition information meets the grammar rule in the scene, and the degree of confidence of the first recognition information is determined to be low when the first recognition information does not meet the grammar rule in the scene.

2. The method of claim 1, wherein prompting the first user to confirm the first recognition information according to the degree of confidence of the first recognition information comprises:

if the degree of confidence of the first recognition information is lower than a first preset threshold, displaying the first recognition information for the first user to confirm;

or

if the degree of confidence of at least one key word in the first recognition information is lower than a second preset threshold, displaying the key word for the first user to confirm.

3. The method of claim 1, wherein extracting, according to second information which is fed back by a second user, rectification information corresponding to the second information comprises:

conducting key word similarity matching on the second information which is fed back by the second user with the confirmed first recognition information, and extracting a matched key word to serve as the rectification information.

4. The method of claim 1, wherein extracting, according to second information which is fed back by a second user, rectification information corresponding to the second information comprises:

extracting the rectification information corresponding to the second information from the confirmed first recognition information according to a key word and/or a language rule in a current dialogue scene.

5. The method of claim 1, further comprising:

determining the current dialogue scene according to the first recognition information, or recognizing the current dialogue scene according to a key word in historical recognition information which has been confirmed by the first user.

12. The method of claim 4, further comprising:

determining the current dialogue scene according to the first recognition information, or recognizing the current dialogue scene according to a key word in historical recognition information which has been confirmed by the first user.

6. An on-line voice translation device, comprising:

a first recognition information module, configured to conduct voice recognition on first voice information input by a first user, so as to obtain first recognition information;

a confirmation prompt module, configured to prompt the first user to confirm the first recognition information;

an information translation module, configured to translate the confirmed first recognition information to obtain and output first translation information;

a rectification information extraction module, configured to extract, according to second information which is fed back by a second user, rectification information corresponding to the second information; and

an information correction module, configured to correct the first translation information according to the rectification information and output the corrected translation information,

wherein the confirmation prompt module comprises;

a degree of confidence determination module, configured to determine a degree of confidence of the first recognition information according to a grammar rule in a current dialogue scene, wherein the degree of confidence of the first recognition information is determined to be high when the first recognition information meets the grammar rule in the scene, and the degree of confidence of the first recognition information is determined to be low when the first recognition information does not meet the grammar rule in the scene; and

a user confirmation module, configured to prompt the first user to confirm the first recognition information according to the degree of confidence of the first recognition information.

7. The device of claim 6, wherein the confirmation prompt module is specifically configured to:

if the degree of confidence of the first recognition information is lower than a first preset threshold, display the first recognition information for the first user to confirm;

or

if the degree of confidence of at least one key word in the first recognition information is lower than a second preset threshold, display the key word for the first user to confirm.

8. The device of claim 6, wherein the rectification information extraction module comprises:

a key word matching sub-module, configured to conduct key word similarity matching on the second information which is fed back by the second user with the confirmed first recognition information, and extract a matched key word to serve as the rectification information.

9. The device of claim 6, wherein the rectification information extraction module comprises:

a language rule sub-module, configured to extract the rectification information corresponding to the second information from the confirmed first recognition information according to a key word and/or a language rule in a current dialogue scene.

10. The device of claim 6, further comprising:

a dialogue scene determination module, configured to determine the current dialogue scene according to the first recognition information, or recognize the current dialogue scene according to a key word in historical recognition information which has been confirmed by the first user.

13. The device of claim 9, further comprising:

a dialogue scene determination module, configured to determine the current dialogue scene according to the first recognition information, or recognize the current dialogue scene according to a key word in historical recognition information which has been confirmed by the first user.

11. A non-transitory computer readable medium comprising computer executable instructions, the computer executable instructions, when being executed by a computer processor, executing an on-line voice translation method, wherein the method comprises:

conducting voice recognition on first voice information input by a first user, so as to obtain first recognition information;

determining a degree of confidence of the first recognition information;

prompting the first user to confirm the first recognition information according to the degree of confidence of the first recognition information;

translating the confirmed first recognition information to obtain and output first translation information;

extracting, according to second information which is fed back by a second user, rectification information corresponding to the second information; and

correcting the first translation information according to the rectification information and outputting the corrected translation information,

wherein determining the degree of confidence of the first recognition information comprises;

determining the degree of confidence of the first recognition information according to a grammar rule in a current dialogue scene, wherein the degree of confidence of the first recognition information is determined to be high when the first recognition information meets the grammar rule in the scene, and the degree of confidence of the first recognition information is determined to be low when the first recognition information does not meet the grammar rule in the scene.