Image indexing using color correlograms

• US 6,246,790 B1
• Filed: 12/28/1998
• Issued: 06/12/2001
• Est. Priority Date: 12/29/1997
• Status: Expired due to Term

• Current Assignee: Cornell Research Foundation Incorporated (Cornell University)
• Original Assignee: Cornell Research Foundation Incorporated (Cornell University)
• Inventors: Huang, Jing, Zhu, Wei-Jing, Mitra, Mandar, Kumar, Shanmugasundaram Ravi
• Primary Examiner(s): Tadayon, Bijan
• Assistant Examiner(s): Alavi, Amir
• Application Number:
  US09/221,472
• Time in Patent Office: 897 Days
• Field of Search: 382/162, 382/163, 382/164, 382/165, 382/166, 382/167, 382/168, 382/169, 382/170, 382/171, 382/172, 382/173, 382/232, 382/233, 382/234, 382/235, 382/236
• US Class Current: 382/162
• CPC Class Codes: G06F 16/58 Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 10/56 relating to colour

Abstract

A color correlogram is a three-dimensional table indexed by color and distance between pixels which expresses how the spatial correlation of color changes with distance in a stored image. The color correlogram may be used to distinguish an image from other images in a database. To create a color correlogram, the colors in the image are quantized into m color values, c_i . . . c_m. Also, the distance values kε[d] to be used in the correlogram are determined where [d] is the set of distances between pixels in the image, and where dmax is the maximum distance measurement between pixels in the image. Each entry (i, j, k) in the table is the probability of finding a pixel of color c_i at a selected distance k from a pixel of color c_i. A color autocorrelogram, which is a restricted version of the color correlogram that considers color pairs of the form (i,i) only, may also be used to identify an image.

Claims

1. A computer-implemented method for indexing images, comprising the steps of:

quantizing colors into color values in an image having a plurality of pixels;

selecting a distance value to be used as the distance between pixels to be evaluated for color value;

analyzing said image according to said color values and said selected distance value;

determining in response to the analyzing step a probability of finding a pixel of a particular color value at said distance value from a selected pixel of a selected color value; and

entering said probability into a color correlogram

whereby the image is represented by the color correlogram for the purpose of indexing the image.

2. The method of claim 1 wherein said analyzing step further comprises the step of counting all the pixels at said distance value in a radius of said selected distance value around said selected pixel.

3. The method of claim 1 wherein said analyzing step comprises the step of counting the number of pixels in a positive horizontal and vertical direction at a distance equal to said selected distance value from said selected pixel.

4. The method of claim 1 wherein said particular color value and said selected color value are the same.

5. The method of claim 1 further comprising the steps of:

selecting a plurality of distance values; and

performing said analyzing step, said determining step, and said entering step using said plurality of distance values.

6. The method of claim 5 wherein the provided corellogram is a banded correlogram and the method further comprises the steps of:

summing, for each color pair, the probability values obtained using the plurality of distance values to obtain a probability sum for each color pair; and

,

entering the probability sum into the banded correlogram.

7. The method of claim 6 wherein the particular color value and the selected color value of each said color pair are the same.

8. The method of claim 1 wherein the provided correlogram is an edge correlogram and the method further comprises the steps of:

segmenting the color values into edge color values and non-edge color values; and

,

determining whether a particular pixel is part of an edge.

9. The method of claim 1 further comprising the steps of:

storing said color correlogram in a database.

10. The method of claim 9 further comprising the steps of:

determining a query color correlogram of a query image; and

comparing the query color correlogram with stored color correlograms in the database to determine a similarity between a stored image and the query image.

11. The method of claim 10 further comprising the step of presenting, as a result of said comparing step, a plurality of stored images which are determined to be similar to said query image.

12. A system for indexing images, comprising:

means for quantizing colors into color values in an image having a plurality of pixels;

means for selecting a distance value to be used as the distance between pixels to be evaluated for color value;

means for analyzing said image according to said color values and said distance value;

means for determining, in response to said analyzing means, a probability of finding a pixel of a particular color value at said distance value from a selected pixel of a selected color value; and

means for entering the probability into a color correlogram,

whereby the image is represented by the color correlogram for the purpose of indexing the image.

13. The system for indexing images of claim 12 wherein said means for analyzing further comprises a means for counting all the pixels at said distance value in a radius of said selected distance value around said selected pixel.

14. The system for indexing images of claim 12 wherein said means for analyzing further comprises a means for counting the number of pixels in a positive horizontal and vertical direction at a distance equal to said selected distance value from said selected pixel.

15. The system for indexing images of claim 12 wherein said particular color value and said selected color value are the same.

16. The system for indexing images of claim 12 further comprising:

means for selecting a plurality of distance values;

means for analyzing said image according to said color values and said plurality of distance values; and

means for determining, in response to said analyzing means, a probability of finding a pixel of a particular color value for each of said plurality of distance values from a selected pixel of a selected color value.

17. The system for indexing images of claim 12 further comprising:

a database for storing said correlogram.

18. The system for indexing images of claim 17 further comprising:

means for determining a query color correlogram of a query image; and

means for comparing the query color correlogram with stored color correlograms in the database to determine a similarity between a stored image and the query image.