Granted Patent B2

US 7,116,443 · App. 10/137,677 · Granted Oct 3, 2006

Abstract

Claims (63)

1. A method for classifying a pixel between neutral and color classes for use in color image processing, comprising the following steps:

a) receiving an input image represented in a first color space;

b) converting the input image to a second color space wherein one coordinate of the second color space represents lightness, the second color space being an L*a*b* color space;

c) selecting a pixel in the second color space representation of the input image to be classified;

d) computing second color space feature vectors associated with the selected pixel including,

d 1 ) selecting a pixel window associated with the selected pixel,

d 2 ) computing a first feature vector (V 1 ) based on the selected pixel, computed using the equation:

V 1 =L*(x,y)/255,

wherein (x,y) are coordinates for the selected pixel within the selected pixel window,

d 3 ) computing a second feature vector (V 2 ) based on the selected pixel window, and

d 4 ) computing a third feature vector (V 3 ) based on the selected pixel window; and

e) classifying the selected pixel between neutral and color classes based on the values computed for the second color space feature vectors.

2. The method as set forth in claim 1 , wherein the first color space is an RGB color space.

3. The method as set forth in claim 1 , further comprising the following step between steps a) and b):

f) processing the input image using a smoothing filter to create a smoothed input image.

4. The method as set forth in claim 3 , step f) further comprising the following steps:

g) selecting a pixel in the input image to be filtered;

h) selecting a pixel window associated with the pixel selected to be filtered;

i) computing a smoothed value for the pixel selected to be filtered based on the average value of the pixels in the selected pixel window;

j) determining if any additional pixels of the input image have not been filtered; and

k) if any additional pixels to be filtered have not been filtered, selecting a next pixel in the input image to be filtered and repeating steps h) through j).

5. The method as set forth in claim 1 , wherein a tree classifier is used to classify the selected pixel in step e).

6. The method as set forth in claim 1 , wherein the selected pixel window in step f) is a 3×3 pixel window.

7. The method as set forth in claim 1 , wherein a neural network is used to classify the selected pixel in step e).

8. The method as set forth in claim 1 , wherein the second feature vector (V 2 ) in step h) is computed using the following equation:

V2=( a *( x− 1, y− 1)+ a *( x,y− 1)+ a * ( x+ 1, y− 1)+ a *( x− 1, y )+ a *( x,y )+ a *( x+ 1, y )+ a *( x − 1, y+ 1)+ a *( x,y+ 1)+ a *( x+ 1, y+ 1))/255,

wherein (x,y) are coordinates for the selected pixel within the selected pixel window.

9. The method as set forth in claim 1 , wherein the third feature vector (V 3 ) in step i) is computed using the following equation:

V 3 =( b *( x− 1, y− 1)+ b *( x,y− 1) + b *( x+ 1, y− 1)+ b *( x− 1, y )+ b *( x,y )+ b * ( x+ 1, y )+ b *( x− 1, y )+ b *( x,y+ 1)+ b *( x+ 1, y 1))/255,

wherein (x,y) are coordinates for the selected pixel within the selected pixel window.

10. A method for neutral pixel detection for use in color image processing, comprising the following steps:

a) receiving an input image represented in a first color space;

b) converting the input image to a second color space wherein one coordinate of the second color space represents lightness, the second color space being an L*a*b* color space;

c) selecting a pixel in the second color space representation of the input image;

d) computing second color space feature vectors associated with the selected pixel including,

d 1 ) selecting a pixel window associated with the selected pixel,

d 2 ) computing a first feature vector (V 1 ) based on the selected pixel, computed using the eguation:

V 1 = L *( x,y )/255,

wherein (x,y) are coordinates for the selected pixel within the selected pixel window,

d 3 ) computing a second feature vector (V 2 ) based on the selected pixel window, and

d 4 ) computing a third feature vector (V 3 ) based on the selected pixel window; and

e) detecting if the selected pixel is neutral based on the values computed for the second color space feature vectors.

11. The method as set forth in claim 10 , further comprising the following step between steps a) and b):

f) processing the input image using a smoothing filter to created a smoothed input image.

12. The method as set forth in claim 11 , wherein the selected pixel window in step f) is a 3×3 pixel window.

13. A method for classifying pixels of an input image between neutral and color classes for use in color image processing, comprising the following steps:

a) receiving a first strip of an input image represented in a first color space wherein the strip is comprised of a plurality of scanlines;

b) converting the input image strip to a second color space wherein one coordinate of the second color space represents lightness;

c) selecting a first pixel in the second color space representation of the input image strip to be classified;

d) computing second color space feature vectors associated with the selected pixel;

e) classifying the selected pixel between neutral and color classes based on the values computed for the second color space feature vectors;

f) determining if any additional pixels of the input image strip to be classified have not been classified;

g) if any additional pixels to be classified have not been classified, selecting a next pixel in the second color space representation of the input image strip to be classified and repeating steps d) through f);

h) determining if any additional scanlines in the input image have not been received in an input image strip; and

i) if any additional scanlines have not been received in an input image strip, appending one or more next scanlines in the input image to the strip while ousting one or more corresponding scanlines from the strip in a first-in-first-out fashion and repeating steps b) through g).

14. The method as set forth in claim 13 , further comprising the following step between steps a) and b):

h) processing the input image strip using a smoothing filter to create a smoothed input image strip.

15. The method as set forth in claim 13 , step d) further comprising the following steps:

h) selecting a pixel window associated with the selected pixel;

i) computing a first feature vector (V 1 ) based on the selected pixel;

j) computing a second feature vector (V 2 ) based on the selected pixel window; and

k) computing a third feature vector (V 3 ) based on the selected pixel window.

16. The method as set forth in claim 15 , wherein the second color space is an L*a*b* color space wherein the selected pixel window in step h) is a 3×3 pixel window.

Assignments (8)