METHOD AND SYSTEM FOR PROVIDING ACCESS TO INFORMATION OF POTENTIAL INTEREST TO A USER

• US 20120310628A1
• Filed: 05/30/2012
• Published: 12/06/2012
• Est. Priority Date: 04/25/2007
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: RATHOD, Priyang, NGUYEN, Phuong, MESSER, Alan
• Granted Patent: US 9,286,385 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 16/7844 using original textual content or text extracted from visual content or transcript of audio data
• CPC Class Codes: G06F 16/7867 using information manually generated, e.g. tags, keywords, comments, title and artist information, manually generated time, location and usage information, user ratings

Abstract

The present invention provides a method and system for providing access to information of potential interest to a user. Closed-caption information is analyzed to find related information on the Internet. User interactions with a TV which receives programming including closed-caption information are monitored to determine user interests or topics.

Claims

1. A method for extracting a sentence from on incoming stream of text corresponding to a program, the method comprising:

retrieving end-of-sentence punctuation marks for a language identified for the incoming stream of text;

locating punctuation marks in the incoming stream of text that match one of more of the retrieved end-of-sentence punctuation marks;

comparing characters around the located punctuation marks to a list of word-punctuation pairs for the identified language to determine when a located punctuation mark is a valid end-of-sentence punctuation marks as opposed to an invalid one not to be considered an end-of-sentence punctuation mark despite its presence in the retrieved end-of-sentence punctuation marks for the identified language; and

for any located valid punctuation marks, identifying a group of words between located valid punctuation marks as sentences.

2. The method of claim 1, wherein the identified language is an estimated language based on a closed caption descriptor in the incoming stream of text.

3. The method of claim 1, wherein the identified language is an estimated language based on a previously identified language for a stream of text earlier in the program.

4. The method of claim 1, wherein the identified sentences are used to aid in a natural language process to tag topics in the incoming stream of text.

5. A method for identifying a language of an incoming stream of text corresponding to a program, the method comprising:

comparing the incoming stream of text against one or more character sets, each of the character sets identifying characters used in a different language;

identifying stop words in the incoming stream of text and comparing the identified stop words with stop words corresponding to one or more languages; and

identifying a particular language for the incoming stream of text based on a marched character set and identified stop words corresponding to the particular language.

6. The method of claim 5, further comprising:

locating a language descriptor in the incoming stream of text, and wherein the identifying a particular language is also based on the language descriptor.

7. The method of claim 5, further comprising maintaining a confidence table that tracks the likelihood of each possible language being the current language of the program.

8. The method of claim 5, further comprising:

identifying a sentence for the incoming stream of text by comparing punctuation marks in the incoming stream of text with character sets for different languages.

9. A method for validating a topic extracted from a stream of text corresponding to a program, the method comprising:

locally validating the topic by comparing the topic against one or more local word lists; and

remotely validating the topic by submitting the topic as a query to an Internet search engine and comparing the number of results received from the Internet search engine to a predefined threshold.

10. The method of claim 9, wherein the one or more local word lists include a list of names.

11. The method of claim 9, wherein the one or more local word lists include a list of locations.

12. The method of claim 9, wherein the remotely validating further comprising determining if the topic can be submitted to a particular widget.

13. An apparatus comprising:

a closed-caption decoder configured to decode a raw closed caption stream for a program and produce closed caption text;

a language detection module configured to determine a language for the closed caption text;

a sentence detection module configured to determine sentences within the closed caption text;

a tagger configured to tag keywords based in the closed caption text based on the determined language and based on determined sentences;

a topic extractor configured to extract topics based on the tagged keywords; and

a validation module configured to validate the extracted topics.

14. The apparatus of claim 13, wherein the tagger includes natural language processing functions.

15. The apparatus of claim 13, wherein the topic extractor uses different rules for different genres of program.

16. The apparatus of claim 15, wherein the rules are stored in a rule library.

17. The apparatus of claim 16, further comprising a memory storing a genre-rule map providing a mapping between genres and rules.

18. The apparatus of claim 13, wherein the validation module includes a local validation module and a remote validation module.

19. The apparatus of claim 13, wherein the language detection module receives character sets and stop words as input.

20. The apparatus of claim 13, wherein the remote validation module is configured to submit extracted topics to an Internet search engine.