IMAGE SEARCH APPARATUS AND METHOD, AND COMPUTER READABLE MEMORY

• US 20020181783A1
• Filed: 04/26/1999
• Published: 12/05/2002
• Est. Priority Date: 04/27/1998
• Status: Active Grant

• Current Assignee: Canon Ayutthaya Limited (Canon Inc.)
• Original Assignee: Canon Ayutthaya Limited (Canon Inc.)
• Inventors: SHIIYAMA, HIROTAKA
• Granted Patent: US 6,567,551 B2
• Time in Patent Office: Days
• US Class Current: 382/217
• CPC Class Codes: G06F 16/583 using metadata automatically derived from the content
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

Image data and their image feature amounts are managed in correspondence with each other using an image storage unit and an information management DB. A desired one of query conditions each of which is defined by image feature amounts set with ambiguity levels is designated by an input unit. A matching unit matches the image feature amounts that define the designated query condition, with corresponding managed image data. Image data as search results are output to a display on the basis of the matching results.

Claims

1. An image search apparatus for managing image data, and searching the managed image data for desired image data, comprising:

management means for managing image data and image feature amounts thereof in correspondence with each other;

designation means for designating a desired one of query conditions each of which is defined by image feature amounts set with ambiguity levels;

matching means for matching the image feature amounts that define the query condition designated by said designation means, with corresponding image feature amounts of image data managed by said management means; and

output means for outputting image data as search results on the basis of matching results of said matching means.

2. The apparatus according to claim 1, wherein said designation means displays, as the query conditions, character strings which express contents of the ambiguity levels set for the image feature amounts.

3. The apparatus according to claim 1, wherein said matching means matches the image feature amounts which define the query condition with the corresponding image feature amounts of the image data using an evaluation formula.

4. The apparatus according to claim 3, wherein the query condition changes the evaluation formula or coefficients used in the evaluation formula.

5. The apparatus according to claim 1, wherein said output means sorts and outputs image data as search results on the basis of coincidence levels of the image feature amounts obtained as matching results of said matching means.

6. The apparatus according to claim 1, wherein said matching means comprises extraction means for extracting in advance image feature amounts of image data, which have a coincidence level not less than a predetermined value, in units of image feature amounts that define the query condition, and

said matching means matches the image feature amounts of image data extracted by said extraction means with the image feature amounts which define the query condition.

7. The apparatus according to claim 6, wherein a minimum one of coincidence levels of image data extracted by said extraction means is set as a coincidence level of image data which are not extracted by said extraction means.

8. An image search method for managing image data, and searching the managed image data for desired image data, comprising:

the management step of managing image data and image feature amounts thereof in a storage medium in correspondence with each other;

the designation step of designating a desired one of query conditions each of which is defined by image feature amounts set with ambiguity levels;

the matching step of matching the image feature amounts that define the query condition designated in the designation step, with corresponding image feature amounts of image data managed in the storage medium; and

the output step of outputting image data as search results on the basis of matching results in the matching step.

9. The method according to claim 8, wherein the designation step includes the step of displaying, as the query conditions, character strings which express contents of the ambiguity levels set for the image feature amounts.

10. The method according to claim 8, wherein the matching step includes the step of matching the image feature amounts which define the query condition with the corresponding image feature amounts of the image data using an evaluation formula.

11. The method according to claim 10, wherein the query condition changes the evaluation formula or coefficients used in the evaluation formula.

12. The method according to claim 8, wherein the output step includes the step of sorting and outputting image data as search results on the basis of coincidence levels of the image feature amounts obtained as matching results in the matching step.

13. The method according to claim 8, wherein the matching step comprises the extraction step of extracting in advance image feature amounts of image data, which have a coincidence level not less than a predetermined value, in units of image feature amounts that define the query condition, and

the matching step includes the step of matching the image feature amounts of image data extracted in the extraction step with the image feature amounts which define the query condition.

14. The method according to claim 13, wherein a minimum one of coincidence levels of image data extracted in the extraction step is set as a coincidence level of image data which are not extracted in the extraction step.

17. The apparatus according to claim 16, wherein said designation means has a user interface for designating numerical values corresponding to the interest levels, and

said matching means converts the interest levels designated via said user interface into ambiguity levels upon matching, and matches the attribute information and image feature amounts that define the query condition, with the corresponding attribute information and image feature amounts of the image data, using an evaluation formula based on the ambiguity levels.

18. The apparatus according to claim 16, wherein the interest level is a weight corresponding to importance levels to be attached to the attribute information and image feature amounts, and increases ambiguity of matching by said matching means as the interest level assumes a smaller value.

19. The apparatus according to claim 16, wherein said output means sorts and outputs image data as search results on the basis of coincidence levels of the image feature amounts obtained as matching results of said matching means.

20. The apparatus according to claim 19, wherein the coincidence level is a normalized coincidence level obtained by dividing by a sum total of the interest levels of the attribute information and image feature amounts.

21. The apparatus according to claim 16, wherein said matching means comprises extraction means for extracting in advance image feature amounts of image data, which have a coincidence level not less than a predetermined value, in units of image feature amounts that define the query condition, and image feature amounts obtained by numerically expressing the attribute information, and

said matching means matches the image feature amounts of image data extracted by said extraction means, with the image feature amounts which define the query condition.

22. The apparatus according to claim 21, wherein a minimum one of coincidence levels of image data extracted by said extraction means is set as a coincidence level of image data which are not extracted by said extraction means.

24. The method according to claim 23, wherein the designation step has a user interface for designating numerical values corresponding to the interest levels, and

the matching step includes the step of converting the interest levels designated via said user interface into ambiguity levels upon matching, and matching the attribute information and image feature amounts that define the query condition, with the corresponding attribute information and image feature amounts of the image data, using an evaluation formula based on the ambiguity levels.

25. The method according to claim 23, wherein the interest level is a weight corresponding to importance levels to be attached to the attribute information and image feature amounts, and increases ambiguity of matching in the matching step as the interest level assumes a smaller value.

26. The method according to claim 23, wherein the output step includes the step of sorting and outputting image data as search results on the basis of coincidence levels of the image feature amounts obtained as matching results in the matching step.

27. The method according to claim 26, wherein the coincidence level is a normalized coincidence level obtained by dividing by a sum total of the interest levels of the attribute information and image feature amounts.

28. The method according to claim 23, wherein the matching step comprises the extraction step of extracting in advance image feature amounts of image data, which have a coincidence level not less than a predetermined value, in units of image feature amounts that define the query condition, and image feature amounts obtained by numerically expressing the attribute information, and

the matching step includes the step of matching the image feature amounts of image data extracted in the extraction step with the image feature amounts which define the query condition.

29. The method according to claim 28, wherein a minimum one of coincidence levels of image data extracted in the extraction step is set as a coincidence level of image data which are not extracted in the extraction step.

15. A computer readable memory that stores program codes of an image search process for managing image data, and searching the managed image data for desired image data, comprising:

a program code of the management step of managing image data and image feature amounts thereof in a storage medium in correspondence with each other;

a program code of the designation step of designating a desired one of query conditions each of which is defined by image feature amounts set with ambiguity levels;

a program code of the matching step of matching the image feature amounts that define the query condition designated in the designation step, with corresponding image feature amounts of image data managed in the storage medium; and

a program code of the output step of outputting image data as search results on the basis of matching results in the matching step.

16. An image search apparatus for managing image data, and searching the managed image data for desired image data, comprising:

management means for managing image data, and attribute information and image feature amounts thereof in correspondence with each other;

designation means for designating the attribute information and image feature amounts, and interest levels for the attribute information and image feature amounts as a query condition used in a search of image data;

matching means for matching the attribute information and image feature amounts that define the query condition designated by said designation means, with corresponding attribute information and image feature amounts of image data managed by said management means, on the basis of the interest levels; and

output means for outputting image data as search results on the basis of matching results of said matching means.

23. An image search method for managing image data, and searching the managed image data for desired image data, comprising:

the management step of managing image data, and attribute information and image feature amounts thereof in a storage medium in correspondence with each other;

the designation step of designating the attribute information and image feature amounts, and interest levels for the attribute information and image feature amounts as a query condition used in a search of image data;

the matching step of matching the attribute information and image feature amounts that define the query condition designated in the designation step, with corresponding attribute information and image feature amounts of image data managed in the storage medium in the management step, on the basis of the interest levels; and

the output step of outputting image data as search results on the basis of matching results in the matching step.

30. A computer readable memory that stores program codes of an image search process for managing image data, and searching the managed image data for desired image data, comprising:

a program code of the management step of managing image data, and attribute information and image feature amounts thereof in a storage medium in correspondence with each other;

a program code of the designation step of designating the attribute information and image feature amounts, and interest levels for the attribute information and image feature amounts as a query condition used in a search of image data;

a program code of the matching step of matching the attribute information and image feature amounts that define the query condition designated in the designation step, with corresponding attribute information and image feature amounts of image data managed in the storage medium in the management step, on the basis of the interest levels; and

a program code of the output step of outputting image data as search results on the basis of matching results in the matching step.