MULTIMEDIA OBJECT RETRIEVAL FROM NATURAL LANGUAGE QUERIES

• US 20110066630A1
• Filed: 09/11/2009
• Published: 03/17/2011
• Est. Priority Date: 09/11/2009
• Status: Active Grant

• Current Assignee: Monument Peak Ventures, LLC (Dominion Harbor Enterprises, LLC)
• Original Assignee: Eastman Kodak Company
• Inventors: Balduccini, Marcello, Girotto, Sara
• Granted Patent: US 8,161,063 B2
• Time in Patent Office: Days
• US Class Current: 707/766
• CPC Class Codes: G06F 16/43 Querying

Abstract

A method for identifying multi-media assets from a multi-media collection, and the method implemented at least in part by a data processing system and comprising the steps of: receiving a search query, the search query, or a derivative thereof, including a possessive phrase, wherein the possessive phrase includes a possessor and a possession category; identifying a denoted object referred to by the possessive phrase using a predefined relationship database that establishes social relationships between objects; identifying one or more additional possessors of the denoted object using the predefined relationship database; generating one or more additional possessive phrases for the denoted objects relating the identified additional possessors to the denoted object; and identifying one or more multi-media assets from the multi-media collection containing the denoted object responsive to the one or more additional possessive phrases.

Claims

1. A method for identifying multi-media assets from a multi-media collection, and the method implemented at least in part by a data processing system and comprising the steps of:

receiving a search query, the search query, or a derivative thereof, including a possessive phrase, wherein the possessive phrase includes a possessor and a possession category;

identifying a denoted object referred to by the possessive phrase using a predefined relationship database that establishes social relationships between objects;

identifying one or more additional possessors of the denoted object using the predefined relationship database;

generating one or more additional possessive phrases for the denoted objects relating the identified additional possessors to the denoted object; and

identifying one or more multi-media assets from the multi-media collection containing the denoted object responsive to the one or more additional possessive phrases.

2. The method of claim 1, further comprising the step of pre-processing the search query to generate the possessive phrase, the pre-processed search query being a derivative of the search query.

3. The method of claim 2, wherein the search query includes an ambiguous possessive phrase, and the pre-processing transforms the ambiguous possessive phrase into a plurality of unambiguous possessive phrases, the generated possessive phrase being one of the plurality of unambiguous possessive phrases.

4. The method of claim 2, wherein the step generating the possessive phrase includes generating the possession category.

5. The method of claim 4, wherein the search query includes a first possession category, and wherein the pre-processing generates a second possession category that is a synonym of the first possession category, and wherein the generated possessive phrase includes the second possession category.

6. The method of claim 1, wherein the step of identifying the denoted object comprises the step of searching the relationship database to identify objects linked to the possessor via the possession category.

7. The method of claim 1, wherein the step of identifying the one or more additional possessors includes searching the relationship database to identify additional possessors linked to the denoted object via respective possession categories.

8. The method of claim 7, wherein each additional possessive phrase is identified at least by combining one of the identified additional possessors with its respective possession category.

9. The method of claim 1, wherein the step of identifying the multi-media assets includes the step of comparing the additional possessive phrases to metadata associated with multi-media assets in the multi-media collection.

10. The method of claim 9 wherein the metadata associated with the multi-media assets includes user-supplied captions.

11. The method of claim 1, wherein the possession category is father, mother, grandmother, grandfather, sister, brother, husband, wife, son, or daughter.

12. The method of claim 1 wherein the predefined relationship database is formed, at least in part, by importing information from a family tree database created using a genealogy software application.

13. The method of claim 1 wherein the one or more multi-media assets are identified, at least in part, using an answer set program.