Method For Red-eye Detection

• US 20110142337A1
• Filed: 08/01/2008
• Published: 06/16/2011
• Est. Priority Date: 08/01/2008
• Status: Active Grant

• Current Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Deonarine, Jonathan, Gaubatz, Matthew
• Granted Patent: US 8,396,261 B2
• Time in Patent Office: Days
• US Class Current: 382/165
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/20092 Interactive image processing based on input by user
• CPC Class Codes: G06T 2207/30216 Redeye defect
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/136 involving thresholding
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 40/19 Sensors therefor

Abstract

A method involving generating a pixel map of a digital image, scanning the pixel map and labeling a found red pixel, where each row of the pixel map is scanned for a red pixel, where a coordinate of each labeled red pixel is stored and compared to surrounding red pixels, arranging found labeled red pixels with previously found labeled red pixels that have a corresponding y-axis coordinate thereby creating a contiguous group, determining if the contiguous group is a red-eye artifact, and creating a list of coordinates that denote red-eye artifacts within the digital image.

Claims

1. A method comprising:

generating a pixel map of a digital image;

scanning the pixel map and labeling a found red pixel;

wherein each row of the pixel map is scanned for a red pixel;

wherein a coordinate of each labeled red pixel is stored and compared to surrounding red pixels;

arranging found labeled red pixels with adjacent previously found labeled red pixels thereby creating a contiguous group;

determining if the contiguous group is a red-eye artifact; and

creating a list of coordinates that denote red-eye artifacts within the digital image.

2. The method of claim 1, further comprising storing coordinates of each labeled red pixel in an array.

3. The method of claim 2, wherein the array is a two-dimensional array.

4. The method of claim 3, wherein a first layer of the array stores the coordinates of red pixels in a present row of the pixel map and wherein a second layer of the array stores previous coordinates of found red pixels in a previously scanned row.

5. The method of claim 4, wherein a coordinate of a found labeled red pixel on first layer of the array is compared to a coordinate of a found labeled red pixel on second layer of the array to determine if the found labeled red pixel and a previous found labeled red pixel have a matching corresponding y-axis coordinate and the previous found pixel resides on the previous row.

6. The method of claim 1, wherein the scanning and storing of found labeled red pixels on the pixel map may be performed by a doubly linked list.

7. The method of claim 5, wherein a new contiguous group is created if the found labeled red pixel has a matching corresponding y-axis coordinate with the previous found labeled red pixel from the previous row and a contiguous group does not already exist.

8. The method of claim 7, wherein an existing contiguous group is expanded to surround all adjacent found labeled red pixels and previously found labeled red pixels.

9. A system for determining the presence of red-eye comprising:

a digital sampler;

wherein a digital image is sampled down by the digital sampler;

a pixel map;

wherein the pixel map is a two dimensional map of the digital image;

a pixel eliminator;

a pixel segmentor;

wherein the pixel segmentor passes over the data using a two dimensional array a single time; and

a group classifier that checks contiguous groups of red pixels to determine if a red-eye condition exist.

10. The system of claim 9, wherein the group classifier discards non red-eye contiguous groups that are outside the user defined boundaries for a red-eye.

11. The system of claim 10, wherein the group classifier compares pixels from the contiguous group with pixels surrounding the pixel group to determine if the contiguous group is a red eye.

12. A method for determining the presence of red eye comprising:

scanning a pixel map and measuring a redness level of each pixel with a redness computation;

wherein the redness computation is completed in one or more stages;

wherein the redness level of each pixel is compared to a defined redness threshold in each stage;

labeling each found red pixel and storing the corresponding coordinates on the pixel map;

arranging found labeled red pixels with previously found labeled red pixels that have a corresponding y-axis coordinate thereby creating a contiguous group;

determining if the contiguous group is a red-eye artifact; and

creating a list of coordinates that denote red-eye artifacts within the digital image.

13. The method of claim 12 wherein the redness computation is performed by integer arithmetic and requires at most one division operation.

14. The method of claim 13 wherein a current pixel that has a redness level above a redness threshold is labeled red.

15. The method of claim 12 wherein the redness computation may use RGB or CieLab components.