USER ACCESSIBLE TISSUE SAMPLE IMAGE DATABASE SYSTEM AND METHOD

• US 20080288450A1
• Filed: 05/14/2007
• Published: 11/20/2008
• Est. Priority Date: 05/14/2007
• Status: Active Grant

• Current Assignee: University of Central Florida Research Foundation Inc. (State University System of Florida)
• Original Assignee: University of Central Florida Research Foundation Inc. (State University System of Florida)
• Inventors: Sugaya, Kiminobu
• Granted Patent: US 7,945,078 B2
• Time in Patent Office: Days
• US Class Current: 707/3
• CPC Class Codes: G06F 16/58 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually

Abstract

Disclosed herein is a user-accessible tissue image database system. The system is searchable using a query image provided by a user. The query image is matched with images in the database based on common visual features. Specifically exemplified is the matching of images based on the presence of a target molecule and/or internal anatomical structure.

Claims

1. A method of querying an image database comprising a plurality of digital images of tissue samples indicating a presence of a target molecule and/or digital images of internal anatomical structures, the method comprising:

uploading a query image with a scale;

removing a portion of background of said query image;

normalizing said query image by comparing scale intensity to an average intensity of said plurality of digital images;

eliminating lightest 20 to 50 percent portion of said query image;

dividing said query image into a plurality of windows;

building percentage color histograms corresponding to said plurality of windows; and

matching said query image to images of said plurality of digital images by comparing histograms.

2. The method of claim 1, wherein said plurality of digital images comprises images of tissue samples treated via in-situ hybridization directed at a target molecule.

3. The method of claim 1, wherein said plurality of digital images comprises images of internal tissue obtained by magnetic resonance imaging.

4. The method of claim 1, wherein said plurality of digital images comprises images of internal tissue obtained by x-ray imaging.

5. The method of claim 1, wherein said query image and scale are in PGM format.

6. The method of claim 1, wherein said removing a portion of background of said query image comprises reducing said query image to 5 to 50 percent of its original size to produce a reduced image;

determining a most common color of said reduced image;

eliminating region of said reduced image comprising said most common color to produce a processed image;

converting said processed image to original size to produce a converted image; and

applying said converted image as a mask over said query image.

7. The method of claim 6, wherein said reduced image is reduced to 5-15 percent of its original size.

8. The method of claim 1, wherein matching said query image comprises comparing histograms of said query image with histograms of said plurality of digital images according to gray values.

9. The method of claim 1, further comprising displaying match images from said plurality of digital images.

10. A user accessible tissue image database system comprising

a computer unit comprising at least one processing module and at least two computer-readable program code modules;

and

a plurality of images stored in said computer unit, wherein said plurality of images are digital images of tissue samples indicating a target molecule;

wherein said at least two computer-readable program code modules comprises

a first computer-readable program code module for causing said computer to upload a query image with a scale;

a second computer-readable program code module for causing said computer to remove a portion of background of said query image;

a third computer readable program code module for causing said computer to normalize said query image by comparing scale intensity to an average intensity of said plurality of digital images;

a fourth computer-readable program code module for causing said computer to eliminate lightest 20 to 50 percent portion of said query image;

a fifth computer-readable program code module for causing said computer to divide said query image into a plurality of windows;

a sixth computer-readable program code module for causing said computer to build percentage color histograms corresponding to said plurality of windows; and

a seventh computer-readable program code module for causing said computer to match said query image to images of said plurality of digital images by comparing histograms.

11. The system of claim 10, wherein said at least two program code modules further comprise an eighth computer-readable program code module for causing said computer to display images matched with said query image.

12. The system of claim 11, wherein said plurality of images are tagged with keywords and wherein said keywords are displayed with images matched with said query image.

13. The system of claim 10, wherein said plurality of digital images comprises images of tissue samples treated via in-situ hybridization directed at a target molecule.

14. The system of claim 10, wherein said plurality of digital images comprises images of internal tissue obtained by magnetic resonance imaging.

15. The system of claim 10, wherein said plurality of digital images comprises images of internal tissue obtained by x-ray imaging.

16. The system of claim 10, wherein said query image and scale are in PGM format.

17. The system of claim 10, wherein said computer unit is connected to a computer network.

18. The system of claim 10, wherein said computer network is the internet.

19. The system of claim 10, wherein eighth computer-readable program code module causes said computer to compare histograms of said query image with histograms of said plurality of digital images according to gray values.

20. A method of querying a user accessible tissue image database system comprising:

providing a computer unit comprising a plurality of digital images stored therein, wherein said plurality of digital images indicate the presence of a target molecule or an internal anatomical structure;

said computer unit connected to a computer network and comprising a plurality of computer-readable program code modules for causing said computer unit to identify images from said plurality of digital images that match another image based on similar visual features;

receiving a query image from a remote user;

comparing said query image with said plurality of digital images;

identifying match images from said plurality of digital images based on similar features to said query image according to predefined criteria; and

displaying said match images to said remote user.

21. The method of claim 20, further comprising displaying keywords to said remote user tagged to match images.

22. The method of claim 20 wherein comparing said query image with said plurality of digital images comprises dividing said query image into a plurality of windows;

building percentage color histograms for said plurality of windows and comparing histograms from said query image with histograms of images of said plurality of digital images according to gray values.