Image database browsing and query using texture analysis

• US 5,933,823 A
• Filed: 03/01/1996
• Issued: 08/03/1999
• Est. Priority Date: 03/01/1996
• Status: Expired due to Term

• Current Assignee: Ricoh Company Limited
• Original Assignee: Ricoh Company Limited, Ricoh Corporation (Ricoh Company Limited)
• Inventors: Hull, Jonathan, Hart, Peter, Cullen, John
• Primary Examiner(s): Feild, Joseph H.
• Assistant Examiner(s): Young, John L.
• Application Number:
  US08/609,641
• Time in Patent Office: 1,250 Days
• Field of Search: 345/333, 345/443, 345/166, 358/453, 358/462, 358/409, 364/725.03, 382/305, 382/215, 707/3, 707/4, 707/5, 707/530, 707/6
• US Class Current: 1/1
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06V 30/40 Document-oriented image-based pattern recognition
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Method and apparatus for querying a document image database based on texture, i.e., analytically discernable patterns in the document images of the database. According to the invention, a document image database could be browsed for documents with a particular texture in a variety of ways. For example, a user could input an example document image with a similar appearance to the desired document. Alternatively, the user could employ a simple interface to define a synthetic document based on selection of a few categories. The synthetic document would then serve as an example for search. Or the user may employ a graphical interface to more precisely define an example for search. Thus, the user'"'"'s knowledge of the general appearance of the desired document or documents provides the basis for the search.

Claims

1. A method for searching a document image database comprising:

developing an example document image;

obtaining image feature information about said example document image;

searching, using said image feature information, the document image database;

displaying search results of said search step;

accepting user input selecting a given search result and;

when requested by said user input, performing further iterative searching by using said selected search result as a new example document image.

2. The method of claim 1 wherein said developing comprises scanning in said example document image.

3. The method of claim 1 wherein said developing comprises accepting user input specifying characteristics of said example document image.

4. The method of claim 1 wherein said developing comprises operating a graphical system wherein a user specifies and arranges components of said example document image.

5. The method of claim 1 wherein said image feature information comprises information about interest points in said example document image.

6. The method of claim 1 wherein said image feature information comprises information about connected components in said example document image.

7. The method of claim 1 wherein said image feature information comprises information about distribution of connected components in said example document image.

8. The method of claim 1 wherein said obtaining comprises determining an image feature vector for said example document image.

9. The method of claim 8 wherein said searching comprises comparing said image feature vector for said example document image to image feature vectors of document images in said database.

10. The method of claim 9 further comprising:

accepting user input specifying a search text string, and

wherein said searching further comprises comparing said search text string to OCR data associated with said document images in said database.

11. The method of claim 1 wherein said displaying comprises displaying one or more document images texturally similar to said example document image.

12. The method of claim 1 wherein said displaying comprises displaying icons representing clusters of images textually similar to said example document image as a search result.

13. The method of claim 1 wherein said document image database comprises document images in encrypted form and image feature information for said encrypted document images.

14. The method of claim 13 wherein said searching comprises searching said image feature information stored in said document image data base, and wherein said displaying step comprises decrypting search results of said searching.

15. The method of claim 1 wherein said document image database comprises document images in compressed form and image feature information for said compressed document images.

16. The method of claim 15 wherein said searching comprises searching said image feature information stored in said document image database, and wherein said displaying comprises decompressing search results of said searching.

17. The method claim 1 wherein said developing comprises accepting user input specifying a category of said document.

18. The method claim 1 wherein said developing comprises accepting user input specifying a number of columns of text of said document.

19. The method claim 1 wherein said developing comprises accepting user input specifying an organization of said document.

20. The method claim 1 wherein said developing comprises:

accepting user input specifying a category of said document;

accepting user input specifying a number of columns of text of said document;

accepting user input specifying an organization of said document; and

developing said example document of said category and having said organization and said number of columns.

21. The method claim 1 wherein said displaying comprises:

comparing image feature information of images of pages of said documents;

identifying pages having a similar appearance based on said image feature information; and

displaying pages having a similar appearance superimposed over one another with offset among them.

22. A method of extracting a feature vector from an electronically stored image comprising:

measuring the size of connected components throughout said image;

identifying interest levels of pixels throughout said image;

identifying connected components and their sizes throughout said image;

storing a histogram of connected component sizes as a first group of elements of said feature vector; and

storing a histogram of interest levels of pixels as a second group of elements of said feature vector.

23. The method of claim 22 further comprising:

storing a histogram of connected component population in vertical sections of said image as a third group of elements of said feature vector.

24. The method of claim 22 further comprising:

storing the number of connected components in a plurality of regions of said image as a fourth group of elements of said feature vector.

25. A document image browsing system comprising:

an electronic storage unit that stores a document image database;

an input device for accepting user input;

a display that displays document images;

a processor unit coupled to said electronic storage device and said display, said processor unit operative to;

develop an example document image;

obtain image feature information about said example document image; and

when requested by said user input, iteratively search said document image database for document images similar to said example document image, wherein said search is performed using said image feature information, and wherein a new search is based on a new example document image that is selected from a search result of a previous search.

26. Software on a storage medium comprising:

software for developing an example document image;

software for obtaining image feature information about said example document image;

software for searching, using said image feature information, the document image database;

software for displaying search results of said searching software;

software for accepting user input selecting a displayed search result; and

software for performing, when requested by said user input, further iterative searching by using said selected search result as a new example document image.

27. The software of claim 26 wherein said developing software comprises software for accepting user input specifying characteristics of said example document image.

28. The software of claim 26 wherein said developing software comprises software for operating a graphical system wherein a user specifies and arranges components of said example document image.

29. The software of claim 26 wherein said obtaining software comprises software for extracting an image feature vector for said example document image.

30. The software of claim 29 wherein said searching software comprises software for comparing said image feature vector for said example document image to image feature vectors of document images in said database.

31. The software of claim 30 further comprising software for accepting user input specifying a search text string and wherein said searching software further comprises software for comparing said search text string to OCR data associated with said document images in said database.

32. The software of claim 26 wherein said developing software comprises software for accepting user input specifying a category of said document.

33. The software of claim 26 wherein said developing software comprises software for accepting user input specifying a number of columns of text of said document.

34. The software of claim 26 wherein said developing software comprises software for accepting user input specifying an organization of said document.

35. The software of claim 26 wherein said developing software comprises:

software for accepting user input specifying a category of said document;

software for accepting user input specifying a number of columns of text of said document;

software for accepting user input specifying an organization of said document; and

software for developing said example document of said category and having said organization and said number of columns.

36. The software of claim 26 wherein said displaying software comprises:

software for comparing image feature information of images of pages of said documents;

software for identifying pages having a similar appearance based on said image feature information; and

software for displaying pages having a similar appearance superimposed over one another with offset among them.