Automatic contrast enhancement

• US 20030053690A1
• Filed: 07/06/2001
• Published: 03/20/2003
• Est. Priority Date: 07/06/2001
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: JASC Software (Corel Holdings LP)
• Inventors: Trifonov, Mikhail Ivanovich, Sharonova, Olga Vadimovna, Zaklika, Krzystof Antoni
• Granted Patent: US 6,826,310 B2
• Time in Patent Office: Days
• US Class Current: 382/168
• CPC Class Codes: G06T 5/40 using histogram techniques
• CPC Class Codes: G06T 5/92 based on global image properties

Abstract

A method of automatic contrast correction comprises providing digital image data, representing the data in the form of a brightness histogram, determining a measure of central tendency for said histogram, adding a shift value to the measure of central tendency to provide an adjusted measure of central tendency, using the both the determined and adjusted measures of central tendency to determine the exponent of a tone reproduction curve, and applying the tone reproduction curve to the brightness values of the image. Additionally, the method may provide further correction by applying to the brightness values of the image an S-shaped tone reproduction curve that does not change a measure of central tendency of the brightness histogram.

Claims

9. A method of contrast correction comprising providing digital image data, representing the data in the form of a brightness histogram, determining a measure of central tendency for said histogram, and applying to brightness values of the image an S-shaped tone reproduction curve that does not change a measure of central tendency of the brightness histogram.

1. A method of automatic contrast correction comprising providing digital image data, representing the data in the form of a brightness histogram, determining a measure of central tendency for said histogram as a determined measure of central tendency, adding a shift value to the measure of central tendency to provide an adjusted measure of central tendency, computing a tone reproduction curve by a ratio containing logarithms of determined and adjusted measures of central tendency, and applying this tone reproduction curve to brightness values of the image.

2. The method of claim 1 wherein computing the exponent of a tone reproduction curve, γ

, is performed by the equation

γ

=ln(L_orig/N)/ln(L_adj/N)

where L_orig represents the determined measure of central tendency, L_adj the adjusted measure of central tendency, and N the maximum image brightness.

3. The method of claim 1 wherein computing the exponent of a tone reproduction curve, γ

, is performed by the equation

γ

=(ln L_orig−

ln N)/(ln L_adj−

ln N)

where L_orig represents the determined measure of central tendency, L_adj the adjusted measure of central tendency, and N the maximum image brightness.

4. The method of claim 1 wherein the shift value is zero for a mean brightness of zero and increases to a maximum with increasing mean brightness before falling to zero at an intermediate mean brightness, and as mean brightness increases beyond the intermediate mean brightness, the shift value decreases, falling to a negative, minimum value before increasing to zero again at the maximum mean brightness that the image can have.

5. The method of claim 1 wherein the tone reproduction curve is given by

L′

_i=N(L_i/N)^1/γ

where L_i is any original brightness value, L′

_i a corresponding enhanced brightness value, N is the largest possible brightness value, and γ

is an exponent.

6. The method of claim 1 wherein the measure of central tendency of the brightness histogram is the mean brightness of the histogram, L_mean.

7. The method of claim 5 wherein the measure of central tendency of the brightness histogram is the mean brightness of the histogram, L_mean.

8. The method of claim 7 wherein the shift is provided by the equation:

11. The method of claim 9 wherein the S-shape tone reproduction curve is a tilde function.

13. The method of claim 11 wherein at least one of the parameters of the tilde function is determined by linear regression on the integral brightness histogram.

14. The method of claim 1 wherein there is also applied an S-shaped tone reproduction curve that does not change a measure of central tendency of the brightness histogram.

15. The method of claim 1 wherein after applying the tone reproduction curve to brightness values of the image, applying to brightness values of the image an S-shaped tone reproduction curve that does not change a measure of central tendency of the brightness histogram.

16. The method of claim 15 wherein a shift value is added to the measure of central tendency to provide an adjusted measure of central tendency, computing the exponent of a tone reproduction curve by means of a ratio containing the logarithms of the determined and adjusted measures of central tendency, and applying to the brightness values of the image a tone reproduction curve derived from the properties of the image.

17. The method of claim 16 wherein the shift value is zero for a mean brightness of zero and increases to a maximum with increasing mean brightness before falling to zero at an intermediate mean brightness, and as mean brightness increases beyond the intermediate mean brightness, the shift value decreases falls to a negative, minimum value before increasing to zero again at the maximum mean brightness that the image can have.

18. The method of claim 16 wherein the tone reproduction curve is given by

L′

_i=N(L_i/N)^1/γ

where L_i is any original brightness value, L′

_i a corresponding enhanced brightness value, N is the largest possible brightness value, and γ

is an exponent.

19. The method of claim 16 wherein the measure of central tendency of the brightness histogram is the mean brightness of the histogram, L_mean.

20. The method of claim 18 wherein the measure of central tendency of the brightness histogram is the mean brightness of the histogram, L_mean.

21. The method of claim 19 wherein the shift is provided by the equation:

22. The method of claim 1 wherein the contrast adjustment is performed in a color space selected from the group consisting of CIE L*u*v*, CIE L*a*b*, YIQ, YUV, YCbCr, YES, YCC, HVC and HSV.

23. The method of claim 16 wherein the contrast adjustment is performed in a color space selected from the group consisting of CIE L*u*v*, CIE L*a*b*, YIQ, YUV, YCbCr, YES, YCC, HVC and HSV.

24. The method of claim 1 wherein the shift is determined from a shift curve with a shape is similar to that shown in FIG. 2.

25. The method of claim 1 wherein the shift is determined from a shift curve with a shape selected from the group consisting of:

a) a curve with two distinct, but contiguous sections, a first section having a positive integration value and second section having a negative integration value;

d) a curve with only positive integration values under the curve; and

e) a curve with only negative integration values under the curve.

26. The method of claim 1 wherein the shift is determined from a shift curve with a shape that has positive integration values in relatively lower brightness areas of the histogram and negative integration values in relatively higher brightness areas of the histogram.

27. A computer comprising hardware and software that can execute the method of claim 1 from memory in the computer.

28. A computer comprising hardware and software that can execute the method of claim 15 from memory in the computer.

29. A computer comprising hardware and software that can execute the method of claim 16 from memory in the computer.

10. A method of automatic contrast correction comprising providing digital image data, representing the data in the form of a brightness histogram, determining a measure of central tendency for said histogram, and applying to brightness values of the image an S-shaped tone reproduction curve that does not change a measure of central tendency of the brightness histogram, wherein the shape of the S-shaped tone reproduction curve is determined by linear regression on the integral brightness histogram.

12. The method of claim 10 wherein the S-shape tone reproduction curve is a tilde function.

27-1. A computer comprising hardware and software that can execute the method of claim 9 from memory in the computer.