Method and system for searching of a video archive

• US 20060031216A1
• Filed: 08/04/2004
• Published: 02/09/2006
• Est. Priority Date: 08/04/2004
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Semple, Bruce P., Stein, Frank L.
• Granted Patent: US 7,487,072 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/71 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

Methods and systems are provided that coordinate the operation of multiple search engines through portions of a multimedia object based on the anticipated accuracy of each search engine. One feature of the present invention provides a content management system that efficiently archives and retrieves multimedia content, for example, based on a confidence level or index. As multimedia content is generated, it is analyzed and cataloged into a set of indexes based on the analysis and confidence level of the analysis. The multimedia content is then stored into one or more databases. When one or more items are requested by a query, the query is analyzed and broken down into components or subqueries. These components or subqueries can then be routed to one or more appropriate search engine as a function of the confidence index. The results from the search engines are then gathered and compiled for presentation to a user.

Claims

1. A method of archiving multimedia data, comprising:

receiving a multimedia data signal having content that progresses based on a timing;

converting the multimedia data signal into a searchable format;

identifying a set of features in the searchable format;

determining a set of confidence values for the searchable format that estimates the accuracy of the features in the searchable format as a function of the timing of the multimedia data signal; and

storing the searchable format into a database that is indexed based on the set of features and the confidence values.

2. The method of claim 1, wherein converting the multimedia data signal comprises:

identifying an event in the multimedia data signal; and

selectively converting the multimedia data signal based on the event.

3. The method of claim 2, wherein identifying the event in the multimedia data signal comprises identifying a time code in the multimedia data signal.

4. The method of claim 2, wherein identifying the event in the multimedia data signal comprises identifying a closed caption control sequence in the multimedia data signal.

5. The method of claim 2, wherein identifying the event in the multimedia data signal comprises identifying a scene change in the multimedia data signal.

6. The method of claim 2, wherein identifying the event in the multimedia data signal comprises detecting a threshold level of audio in the multimedia data signal.

7. The method of claim 2, wherein identifying the event in the multimedia data signal comprises identifying a slate in a video stream of the multimedia data signal.

8. The method of claim 1, wherein converting the multimedia data signal into the searchable format comprises:

receiving at least one of the set of confidence values for a previous portion of the multimedia data signal; and

adjusting the conversion of a current portion of the multimedia data signal based on the at least one confidence value.

9. The method of claim 8, wherein adjusting the conversion of the current portion of the multimedia data signal comprises:

comparing the at least one confidence value to a threshold; and

adjusting the conversion of the current portion of the multimedia data signal based on the comparison of the at least one confidence value to the threshold.

10. The method of claim 1 further comprising:

receiving a query that specifies a set of temporal parameters and content parameters;

parsing the query into a set of sub-queries;

selecting respective search engines for each of the sub-queries based on the time reference and the set of confidence values;

gathering a first set of results from each of the search engines that satisfy their respective sub-queries; and

combining the first set of results into a second set of results based on satisfying the temporal parameters and content parameters of the query.

11. The method of claim 10, wherein selecting the respective search engines for each of the subqueries comprises selecting respective search engines based on a set of location parameters in addition to the temporal parameters and content parameters.

12. The method of claim 10, wherein combining the first set of results into a second set of results comprises identifying results from different multimedia data items that occur within a period of time that is specified in the query.

13. The method of claim 10, wherein receiving the query comprises receiving a query that specifies temporal constraints across a plurality of media types.

14. The method of claim 10, wherein receiving the query comprises receiving a query that specifies an offset of time within a multimedia data item.

15. The method of claim 10, wherein parsing the query into sub-queries comprises parsing the query into sub-queries based on a type of media specified by the content parameters.

16. The method of claim 10, wherein selecting the respective search engines for each of the sub-queries based on the time reference and the set of confidence values comprises selecting a first search engine that searches a first time period of a multimedia item based on the confidence values and selecting a second search engine that searches a second time period of the multimedia item based on the confidence values.

17. The method of claim 10, wherein combining the first set of results into the second of results based on satisfying the temporal parameters and content parameters of the query comprises combining the first set of results into a set of references to multimedia data items that are ranked in terms of their relevance to the temporal parameters and the content parameters.

18. A system for archiving multimedia data, comprising:

means for receiving a multimedia data signal having content that progresses based on a timing;

means for converting the multimedia data signal into a searchable format;

means for identifying a set of features in the searchable format;

means for determining a set of confidence values for the searchable format that estimates the accuracy of the features in the searchable format as a function of the timing of the multimedia data signal; and

means for storing the searchable format into a database that is indexed based on the set of features and the confidence values.

19. A system for archiving multimedia data, comprising:

a detector that detects a multimedia data signal having content that progresses based on a timing;

a converter, coupled to the detector, that converts the multimedia data signal into a searchable format;

an analyzer, coupled to the converter, that identifies a set of features in the searchable format, and determines a set of confidence values for the searchable format, wherein the set of confidence values estimate the accuracy of the features in the searchable format as a function of the timing of the multimedia data signal; and

a set of storage devices that store the searchable format into a database that is indexed based on the set of features and the confidence values.

20. The system of claim 19 further comprising:

a server configured to receive a query that specifies a set of temporal parameters and content parameters, parse the query into a set of sub-queries, select respective search engines for each of the sub-queries based on the time reference and the set of confidence values, gather a first set of results from each of the search engines that satisfy their respective sub-queries, and combine the first set of results into a second set of results based on satisfying the temporal parameters and content parameters of the query.