Information processing apparatus and method, program and recording medium

• US 7,805,389 B2
• Filed: 09/27/2007
• Issued: 09/28/2010
• Est. Priority Date: 10/02/2006
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Kobayashi, Yoshiyuki, Saito, Mari, Kimura, Takaomi
• Primary Examiner(s): HOLMES, MICHAEL B
• Application Number:
  US11/863,055
• Publication Number:
  US 20080082506A1
• Time in Patent Office: 1,097 Days
• Field of Search: 706/45
• US Class Current: 706/45
• CPC Class Codes: G06F 16/61 Indexing; Data structures therefor; Storage structures

Abstract

Herein disclosed an information processing apparatus including converting means and retrieval means, wherein the converting means converts content feature quantities using functions adapted to convert a plurality of feature quantities attached to a plurality of pieces of content so that the distance between pieces of content defined by the plurality of feature quantities coincides with the distance suited for a user-entered similarity relationship between the plurality of pieces of content, the functions being further adapted to map the pieces of content laid out in a feature quantity space defined by the plurality of feature quantities into a new feature quantity space by the conversion of the plurality of feature quantities, and wherein the retrieval means retrieves similar pieces of content based on converted feature quantities.

Claims

1. An information processing apparatus comprising:

converting means for converting content feature quantities using functions adapted to convert a plurality of feature quantities attached to a plurality of pieces of content so that the distance between pieces of content defined by the plurality of feature quantities coincides with the distance suited for a user-entered similarity relationship between the plurality of pieces of content, the functions being further adapted to map the pieces of content laid out in a feature quantity space defined by the plurality of feature quantities into a new feature quantity space by the conversion of the plurality of feature quantities; and

retrieval means for retrieving similar pieces of content based on converted feature quantities.

2. The information processing apparatus according to claim 1 further comprising

extraction means for extracting content feature quantities.

3. The information processing apparatus according to claim 1, wherein

the retrieval means retrieves similar pieces of content by retrieving pieces of content in the same group among those groups of content whose feature quantities have been converted by the functions.

4. The information processing apparatus according to claim 1, wherein

the converting means converts content feature quantities by applying the functions obtained by repeated generation through a genetic method until an evaluation value is no longer updated a predetermined number of generations, the evaluation value being an information criterion calculated from two values, one of which is a value representing the degree of classification based on Euclidean distance between pieces of content considered to be similar and dissimilar in the user-entered similarity relationship between the plurality of pieces of content, and the other of which is the number of free parameters of the functions.

5. The information processing apparatus according to claim 4, wherein

the converting means converts content feature quantities by applying the functions obtained by repeated generation through a genetic method until the evaluation value is no longer updated a predetermined number of generations, the evaluation value being calculated from an FDR (Fisher discriminant ratio) which represents the degree of classification.

6. The information processing apparatus according to claim 1 further comprising

generation means for generating the functions through a generic method, wherein

the converting means converts content feature quantities by applying the generated functions.

7. The information processing apparatus according to claim 6 further comprising

calculation means for calculating an evaluation value, which is an information criterion, from two values, one of which is a value representing the degree of classification based on Euclidean distance between pieces of content considered to be similar and dissimilar in the user-entered similarity relationship between the plurality of pieces of content, and the other of which is the number of free parameters of the functions, and wherein

the generation means generates the functions through a genetic method until the evaluation value is no longer updated a predetermined number of generations.

8. The information processing apparatus according to claim 7, wherein

the calculation means calculates the evaluation value from an FDR which represents the degree of classification.

9. The information processing apparatus according to claim 7 further comprising

acquisition means for acquiring data representing a user-entered similarity relationship between the plurality of pieces of content, wherein

the calculation means calculates an evaluation value using the data representing the user-entered similarity relationship between the plurality of pieces of content.

10. The information processing apparatus according to claim 1 further comprising

clustering means for performing clustering of content based on converted feature quantities.

11. The information processing apparatus according to claim 1 further comprising

identification means for identifying a group to which content belongs from a content-to-cluster distance based on converted feature quantities.

12. An information processing method comprising the steps of converting and retrieving, wherein

the converting step converts content feature quantities using functions adapted to convert a plurality of feature quantities attached to a plurality of pieces of content so that the distance between pieces of content defined by the plurality of feature quantities coincides with the distance suited for a user-entered similarity relationship between the plurality of pieces of content, the functions being further adapted to map the pieces of content laid out in a feature quantity space defined by the plurality of feature quantities into a new feature quantity space by the conversion of the plurality of feature quantities, and wherein

the retrieval step retrieves similar pieces of content based on converted feature quantities.

13. A program causing a computer to execute the steps of converting and retrieving, wherein

the converting step converts content feature quantities using functions adapted to convert a plurality of feature quantities attached to a plurality of pieces of content so that the distance between pieces of content defined by the plurality of feature quantities coincides with the distance suited for a user-entered similarity relationship between the plurality of pieces of content, the functions being further adapted to map the pieces of content laid out in a feature quantity space defined by the plurality of feature quantities into a new feature quantity space by the conversion of the plurality of feature quantities, and wherein

the retrieval step retrieves similar pieces of content based on converted feature quantities.

14. A recording medium for storing a program causing a computer to execute the steps of converting and retrieving, wherein

the converting step converts content feature quantities using functions adapted to convert a plurality of feature quantities attached to a plurality of pieces of content so that the distance between pieces of content defined by the plurality of feature quantities coincides with the distance suited for a user-entered similarity relationship between the plurality of pieces of content, the functions being further adapted to map the pieces of content laid out in a feature quantity space defined by the plurality of feature quantities into a new feature quantity space by the conversion of the plurality of feature quantities, and wherein

the retrieval step retrieves similar pieces of content based on converted feature quantities.

15. An information processing apparatus comprising:

a converting section configured to convert content feature quantities using functions adapted to convert a plurality of feature quantities attached to a plurality of pieces of content so that the distance between pieces of content defined by the plurality of feature quantities coincides with the distance suited for a user-entered similarity relationship between the plurality of pieces of content, the functions being further adapted to map the pieces of content laid out in a feature quantity space defined by the plurality of feature quantities into a new feature quantity space by the conversion of the plurality of feature quantities; and

a retrieval section configured to retrieve similar pieces of content based on converted feature quantities.