Determination of an illuminant of digital color image by segmentation and filtering

• US 20030095704A1
• Filed: 11/22/2002
• Published: 05/22/2003
• Est. Priority Date: 11/22/2001
• Status: Active Grant

• Current Assignee: Monument Peak Ventures, LLC (Dominion Harbor Enterprises, LLC)
• Original Assignee: Eastman Kodak Company
• Inventors: Risson, Valery J.
• Granted Patent: US 7,068,840 B2
• Time in Patent Office: Days
• US Class Current: 382/162
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 10/26 Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
• CPC Class Codes: G06V 10/28 Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: H04N 1/6086 by scene illuminant, i.e. conditions at the time of picture capture, e.g. flash, optical filter used, evening, cloud, daylight, artificial lighting, white point measurement, colour temperature

Abstract

The present invention is in the technical field of analyzing and processing digitized color images. More specifically it relates to a method of determining the chromaticity of an illuminant (white reference) of a color image representing a natural scene of objects or entities. The present invention uses the principle of the constancy of dichromatic color, but it also uses spatial color segmentation of the image and relevant filtering to select the regions of the image to be analyzed, that is to extract the analyzed regions non-compliant with the chromatic model. The invention is applicable for improving the performance of the means of analysis and recognition of the shapes of objects in an image.

Claims

1. An automatic method for determining an illuminant of a digital color image, the method comprising the steps of:

a) segmenting an image into regions of homogeneous color;

b) extracting the regions from the segmented image that are non-compliant with a dichromatic reflection model;

c) for each region not extracted from the image, determining from chromatic coordinates of pixels forming said region, a straight line representing a direction of a chromatic convergence;

d) determining a most probable convergence point of a set of the straight lines obtained in said step c); and

e) determining chromatic coordinates of the most probable convergence point obtained in said step d) and representing an illuminant of the color image.

2. The method according to claim 1, further comprising a preliminary step of pre-filtering the image, just before the step of segmenting the image into regions of homogeneous color.

3. The method according to claim 1, wherein the segmentation of the image divides said image spatially into regions of homogeneous color.

4. The method according to claim 1, wherein the segmentation of the image is performed by applying a watershed to a gradient of the image, said gradient being calculated by using a Euclidean distance in a CIE-Lab space.

5. The method according to claim 3, wherein the regions of homogeneous color are regions of homogeneous chrominance.

6. The method according to claim 3, wherein the regions of homogeneous color are regions of homogeneous luminance.

7. The method according to claim 1, wherein the regions extracted in said step b) are shadow regions of the image.

8. The method according to claim 1, wherein the regions extracted in said step b) are sky regions of the image.

9. The method according to claim 1, wherein the regions extracted in said step b) are achromatic regions of the image.

10. The method according to claim 1, wherein for each non-extracted region of the image, the straight line representing the direction of the chromatic convergence is determined by linear regression from the chromatic coordinates of pixels forming said region.

11. The method according to claim 1, wherein the most probable convergence point is the point of a chromatic diagram through which pass a maximum of straight lines obtained in said step c).

12. The method according to claim 1, wherein the most probable convergence point is the point of a diagram of daytime illuminants through which pass a maximum of straight lines obtained in said step c).