Apparatus and method for analyzing utterance meaning

Patent metadata

Field	Value
Publication number	US-10347243-B2
Application number	US-201615377847-A
Country	US
Kind code	B2
Filing date	Dec 13, 2016
Priority date	Oct 5, 2016
Publication date	Jul 9, 2019
Grant date	Jul 9, 2019

Abstract

Disclosed herein is a method for analyzing an utterance meaning. The method includes collecting a voice signal of an utterer; converting the collected voice signal into information in a text form, extracting a keyword of the text information from the text information, and deriving at least one utterance topic on the basis of the extracted keywords of the text information.

First claim

What is claimed is:

1. A method for analyzing an utterance meaning, the method comprising:
selecting an operation mode from a plurality of operation modes, wherein the selected operation mode is for analyzing a topic and meaning of an event based on the utterance meaning of users in the event;
setting a topic keyword having a range of preset topic keywords;
identifying a main utterance topic based on the topic keyword by
collecting a voice signal of an utterer using a voice sensor,
converting, by a controller, the collected voice signal into a single text information in a text form, wherein the controller includes at least one integrated circuit,
extracting, by the controller, a plurality of keywords of the single text information from the single text information,
deriving, by the controller, the main utterance topic based on the extracted plurality of keywords of the single text information, wherein the main utterance topic is selected based on the range of preset topic keyword of the set topic keyword;

analyzing, by the controller, the utterance meaning based on the derived main utterance topic, wherein analyzing the utterance meaning comprises
collecting a further voice signal of the utterer using the voice sensor,
extracting a further keyword by converting the collected further voice signal into a text form,
deriving a range of derived utterance topics based on the main utterance topic and the further keyword,
deriving a detailed utterance topic within the range of derived utterance topics, and
generating a tree according to the detailed utterance topic; and

displaying content on the generated tree, and for each detailed utterance topic displaying a main content of subsequent utterances in the order of utterances of the utterer in real time.

2. The method according to claim 1 , wherein the main utterance topic is derived based on extracted plurality of keywords of the single text information emerging at a preset reference value or higher among the keywords of the text information.

3. The method according to claim 1 , wherein deriving the main utterance topic further comprises deriving a plurality of utterance topics and then deriving a selected utterance topic based on a selection by a user from the utterance topics derived based on the extracted plurality of keywords.

4. The method according to claim 1 , wherein deriving the main utterance topic further comprises deriving an utterance topic based on a correlation between the extracted plurality of keywords.

5. The method according to claim 1 , wherein deriving the main utterance topic further comprises deriving an utterance topic based on information about the utterer.

6. The method according to claim 1 , wherein collecting the voice signal comprises collecting a voice signal from a plurality of utterers and wherein deriving the main utterance topic comprises deriving an utterance topic based on common information about the utterers.

7. The method according to claim 1 , wherein deriving the main utterance topic further comprises deriving an utterance topic selected at a preset reference value or higher based on a history of selecting utterance topics.

8. The method according to claim 1 , wherein deriving the main utterance topic comprises extracting information about utterance intent, purpose, or target of the utterer based on the extracted plurality of keywords of the single text information, and deriving an utterance topic based on the extracted information.

9. The method according to claim 1 , wherein deriving the main utterance topic comprises deriving an utterance topic based on a result of analyzing the utterance meaning.

10. An apparatus for analyzing an utterance meaning comprising:
a collection unit configured to collect a voice signal of an utterer;
a display; and
a controller implemented using at least on integrated circuit;
a memory storing a program to be executed in the controller, the program comprising instructions to
select an operation mode from a plurality of operation modes, wherein the selected operation mode is for analyzing a topic and meaning of an event based on the utterance meaning of users in the event;
set a topic keyword having a range of preset topic keywords;
identify a main utterance topic based on the topic keyword by
receive a voice signal of the utterer from the voice sensor,
convert the collected voice signal into a single text information in a text form,
extract a plurality of keywords of the single text information from the single text information,
derive the main utterance topic based on the extracted plurality of keywords of the single text information, wherein the main utterance topic is selected based on the range of preset topic keyword of the set topic keyword;

analyze the utterance meaning based on the derived main utterance topic by
receive a further voice signal of the utterer from the voice sensor,
extract a further keyword by converting the collected further voice signal into a text form,
derive a range of derived utterance topics based on the main utterance topic and the further keyword,
derive a detailed utterance topic within the range of derived utterance topics, and
generate a tree according to the detailed utterance topic; and
provide a content on the generated tree to the display, and for each detailed utterance topic provide a main content of subsequent utterances in the order of utterances of the utterer in real time.

11. The apparatus according to claim 10 , wherein the program comprises further instructions to cause the controller to derive the main utterance topic based on the extracted plurality of keywords of the single text information output above a preset reference value among the extracted plurality of keywords of the single text information.

12. The apparatus according to claim 10 , wherein the program comprises further instructions to cause the controller to derive the main utterance topic based on a selection, by a user, from the utterance topics derived based on the extracted plurality of keywords.

13. The apparatus according to claim 10 , wherein the program comprises further instructions to cause the controller to derive the main utterance topic based on a correlation between the extracted plurality of keywords.

14. The apparatus according to claim 10 , wherein the program comprises further instructions to cause the controller to derive the main utterance topic based on information about the utterer.

15. The apparatus according to claim 10 , wherein the voice signal comprises a voice signal of a plurality of utterers, and wherein the controller is configured to derive the main utterance topic based on common information about the utterers.

16. The apparatus according to claim 10 , wherein the program comprises further instructions to cause the controller to derive, as the main utterance topic of the utterer, an utterance topic selected at a preset reference value or higher based on a history of selecting the utterance topic.

17. The apparatus according to claim 10 , wherein the program comprises further instructions to cause the controller to extract information about an utterance intent, purpose, or target of the utterer based on the keywords, and to derive an utterance topic based on the extracted information.

18. The apparatus according to claim 10 , wherein the program comprises further instructions to cause the controller to derive the main utterance topic based on a result of analyzing the utterance meaning.

19. The apparatus according to claim 10 , wherein the

Assignees

• Hyundai Motor Co Ltd

Inventors

• Lee Jia
• Han Jimin
• Park Jung Mi

Classifications

• G10L15/22

  Procedures used during a speech recognition process, e.g. man-machine dialogue · CPC title

• G10L15/1815 Primary

  Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning · CPC title

• G10L15/02

  Feature extraction for speech recognition; Selection of recognition unit · CPC title

• G10L2015/088

  Word spotting · CPC title

• G06F16/68 Primary

  Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually · CPC title