Searching for object images with reduced computation

• US 7,191,164 B2
• Filed: 08/19/2003
• Issued: 03/13/2007
• Est. Priority Date: 08/19/2003
• Status: Expired due to Fees

• Current Assignee: Indian Institute of Technology Delhi, Intel Corporation
• Original Assignee: Indian Institute of Technology Delhi, Intel Corporation
• Inventors: Acharya, Tinku, Mishra, Ranjit K., Ray, Ajoy K.
• Primary Examiner(s): Hirl; Joseph P
• Application Number:
  US10/643,467
• Publication Number:
  US 20050044056A1
• Time in Patent Office: 1,302 Days
• Field of Search: 706/52, 706/46, 706/14, 345/355, 382/118
• US Class Current: 706/52
• CPC Class Codes: G06F 16/532 Query formulation, e.g. graphical querying
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

In one embodiment, the present invention includes a method to obtain a query image and search a database corresponding to object images for a solution set having a maximum similarity to the query image using fuzzy logic. Also in certain embodiments, based on fuzzy logic, the database may be partitioned into multiple sets based on a fuzzy similarity analysis of a measure of the object images to various thresholds.

Claims

1. A method comprising:

partitioning a database corresponding to object images into a first partition and a second partition based on a fuzzy similarity analysis of a measure of the object images to a first threshold;

partitioning each of the first partition and the second partition into at least two portions so that the measure of the object images having a fuzzy similarity more than or equal to a second threshold cluster into a selected one of the at least two portions;

determining a feature set from image content of a query object image;

after partitioning the first partition into the at least two portions. using fuzzy logic to search the database for at least one image similar to the query object image; and

outputting the at least one image similar to the query object image.

2. The method of claim 1 further comprising:

deriving the feature set for each of the object images from contours of at least two views of objects corresponding to each of the object images.

3. The method of claim 1, wherein using the fuzzy logic comprises comparing one object image from each of said first and second partitions with said query object image.

4. The method of claim 3, further comprising:

based on the comparison, obtaining the at least one similar image as a match in the partition that indicates maximum similarity with said query object image.

5. The method of claim 1, further comprising:

forming a similarity matrix for the object images within the database before partitioning the database.

6. A method comprising:

partitioning a database corresponding to object images into a plurality of sets based on fuzzy logic;

obtaining a query image;

after partitioning the database into the plurality of sets, searching the database for a solution set having a maximum similarity to the query image using fuzzy logic, and

outputting at least a portion of the solution set.

7. The method of claim 6, wherein searching the database comprises comparing a single image of each of a plurality of sets within the database to the query image.

8. The method of claim 7, wherein comparing the single image comprises comparing a feature vector of the query image to a corresponding feature vector of the single image.

9. The method of claim 6, further comprising partitioning the database into a plurality of levels, each of the levels corresponding to a similarity threshold.

10. The method of claim 6, wherein outputting a portion of the solution set comprises displaying at least one object image corresponding to the portion of the solution set.

11. An article comprising a machine-readable storage medium containing instructions that if executed enable a system to:

partition a database corresponding to object images into a plurality of sets based on fuzzy logic;

obtain a query image;

after the database is partitioned, search the database for a solution set having a maximum similarity to the query image using the fuzzy logic; and

output at least a portion of the solution set.

12. The article of claim 11, further comprising instructions that if executed enable the system to compare a single image of each of a plurality of sets within the database to the query image.

13. The article of claim 12, further comprising instructions that if executed enable the system to compare a feature vector of the query image to a corresponding feature vector of the single image.

14. A system comprising:

a dynamic random access memory containing instructions that when executed enable the system to partition a database corresponding to object images into a first partition and a second partition based on a fuzzy similarity analysis of a measure of the object images to a first threshold;

to thereafter use fuzzy logic to search the database for at least one image similar to a query object image; and

to output the at least one image similar to query object image; and

a processor coupled to the dynamic random access memory to execute the instructions.

15. The system of claim 14, further comprising instructions that when executed enable the system to derive a feature set for each of the object images from contours of at least two views of objects corresponding to each of the object images.

16. The system of claim 14, further comprising instructions that when executed enable the system to obtain the at least one similar image as a match in the partition that indicates maximum similarity with said query object image.

17. The system of claim 16, further comprising a display coupled to the processor to display the query object image and the at least one similar image.