Image forming system, apparatus and method of discriminative color features extraction thereof

• US 8,331,667 B2
• Filed: 07/29/2008
• Issued: 12/11/2012
• Est. Priority Date: 09/28/2007
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Chen, Maolin, Jeong, Moon Sik
• Primary Examiner(s): Wu, Jingge
• Assistant Examiner(s): Patel, Nirav G
• Application Number:
  US12/216,707
• Publication Number:
  US 20090087088A1
• Time in Patent Office: 1,596 Days
• Field of Search: None
• US Class Current: 382/168
• CPC Class Codes: G06F 18/2321 using statistics or function optimisation, e.g. modelling of probability density functions
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: G06V 10/763 Non-hierarchical techniques, e.g. based on statistics of modelling distributions
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands

Abstract

Provided are an apparatus and method of extracting a discriminative color feature, and an image forming system including: a photographing device to photograph an image of an object; a color feature extracting device receiving the image from the photographing device, extracting a discriminative color feature from the image, generating a final color model of the object, extracting a color blob of the object based on the final color model, performing blob analysis on the extracted color blob, and generating parameters to control a posturing of the photographing device according to the blob analysis; a control device receiving from the color feature extracting device the parameters to control the posturing of the photographing device, and controlling the posturing of the photographing device; a storage device storing the photographed image of the object; and a display device displaying the photographed image of the object.

Claims

1. A color feature extracting device, comprising:

a main color extracting unit determining an object area of an input image corresponding to a size and a location of an object, dividing the object area and a non-object area of the input image into a plurality of rectangles, generating a color histogram for each color channel, and extracting a main color of an object rectangle from the color histogram;

a weight calculating unit calculating a minimum distance between a color histogram bin of the object rectangle and a color histogram bin of a non-object rectangle with respect to the main color, a bin weight of the color histogram of the object rectangle, a color channel weight of the object rectangle, and a weight of the object rectangle to extract a discriminative color feature of the object; and

a final color model generating unit generating a final color model of the object based on a result of re-weighting the color histogram bins of the object rectangle.

2. The color feature extracting device of claim 1, wherein the main color extracting unit comprises:

an area divider determining the object area of the input image, dividing the object area and the non-object area of the input image into the plurality of rectangles;

a histogram calculator generating the color histogram of the object rectangle and of the non-object rectangle and the color histogram of the non-object rectangle using the color channel, each color histogram being divided into a plurality of discrete bins; and

a main color extractor extracting the main color of the object rectangle and of the non-object rectangle based on the generated color histogram of the object rectangle.

3. The color feature extracting unit of claim 1, wherein the weight calculating unit comprises:

a minimum bin distance calculator calculating the minimum distance between the color histogram bin of the object rectangle and the color histogram bin of the non-object rectangle with respect to the main color;

a bin weight calculator calculating the weight of the color histogram bin of the object rectangle based on the calculated minimum distance;

a color channel weight calculator calculating the color channel weight of the object rectangle based on the calculated bin weight;

a rectangle weight calculator calculating the weight of the object rectangle based on the calculated bin weight; and

a bin re-weight calculator re-weighting the color histogram bin of the object rectangle based on the color channel weight and the weight of the object rectangle to extract the discriminative color feature of the object.

4. The color feature extracting device of claim 3, wherein the bin minimum distance calculator calculates the minimum distance between the color histogram bin of the object rectangle and the color histogram bin of the non-object rectangle according to,

5. The color feature extracting device of claim 3, wherein the bin weight calculator calculates the weight of the color histogram bin of the object rectangle according to,

w_r^c(i)=T(D_r^c(i),s)×

|H_r^c(i)−

BG^c(J)|,

where s=sign(H_r^c(i)−

BG^c(J)),T(x,y)=e<sup>x×

k×

y</sup>, and k is a constant.

6. The color feature extracting device of claim 5, wherein the color channel weight calculator calculates the weight of the color channel of the object rectangle according to,

7. The color feature extracting device of claim 6, wherein the rectangle weight calculator calculates the weight of the object rectangle according to,

8. The color feature extracting device of claim 7, wherein the bin re-weight calculator re-weights the color histogram bin of the object rectangle, based on the color channel weight and the weight of the object rectangle, according to,

W_r^c(i)=w_r^c(i)×

W_r×

W^c

and thereby calculates the re-weighted color histogram bin of the object rectangle W_r^c(i).

9. The color feature extracting device of claim 8, wherein the final color model generating unit generates the final color model of the object according to,

N={N^r,N^g,N^b},

where, with respect to the color channel,

10. The color feature extracting device of claim 1, further comprising:

a color blob extracting unit extracting a color blob of the object from a new input image based on the final color model of the object.

11. The color feature extracting device of claim 10, further comprising:

an object locating unit performing blob analysis on the extracted color blob and calculating a centroid and a size of an object in the input image based on the blob analysis to locate and track the object.

12. A method of extracting a discriminative color feature, comprising:

determining an object area of an input image corresponding to a size and a location of an object;

dividing the object area and a non-object area of the input image into a plurality of rectangles, and generating a color histogram for each color channel to extract a main color of an object rectangle from the color histogram;

calculating a minimum distance between a color histogram bin of the object rectangle and a color histogram bin of the non-object rectangle to identify a discriminative color histogram bin of the object rectangle from the histogram bin of the non-object rectangle;

calculating a bin weight of the color histogram of the object rectangle based on the calculated minimum distance;

calculating a color channel weight of the object rectangle and a weight of the object rectangle based on the calculated bin weight of the color histogram of the object rectangle;

re-weighting the color histogram bin of the object rectangle based on the calculated color channel weight and the weight of the object rectangle to extract a discriminative color feature of the object; and

generating a final color model of the object based on the result of re-weighting the color histogram bins of the object rectangle.

13. The method of claim 12, wherein the calculating of the minimum distance calculates the minimum distance between the color histogram bin of the object rectangle and the color histogram bin of the non-object rectangle according to,

14. The method of claim 12, wherein the calculating of the bin weight calculates the weight of the color histogram bin of the object rectangle according to,

w_r^c(i)=T(D_r^c(i),s)×

|H_r^c(i)−

BG^c(J)|,

where s=sign(H_r^c(i)−

BG^c(J)),T(x, y)=e<sup>x×

k×

y</sup>, and k is a constant.

15. The method of claim 14, wherein the color channel weight and the weight of the object rectangle is calculated according to,

16. The method of claim 15, wherein the calculating of the color channel weight and the weight of the object rectangle calculates the weight of the object rectangle according to,

17. The method of claim 16, wherein the color histogram bin of the object rectangle is re-weighted based on the color channel weight and the weight of the object rectangle, according to,

W_r^c(i)=w_r^c(i)×

W_r×

W^c,

and thereby calculates the re-weighted color histogram bin of the object rectangle W_r^c(i).

18. The method of claim 17, wherein the final color model of the object is generated according to,

N={N^r,N^g,N^b},

where, with respect to the color channel,

19. The method of claim 12, further comprising:

extracting a color blob of the object from a new input image based on the final color model of the object.

20. The method of claim 19, further comprising:

performing blob analysis on the extracted color blob and calculating a centroid and a size of an object in the input image based on the blob analysis to locate and track the object.