Black Recalculation for Arbitrary HSL Corrections in CMY Color Space

The present application is related to my copending application, Hill Case No. P-92,1545 titled "HSL CORRECTIONS IN CMY COLOR SPACE".

This patent describes color manipulation for digital image processing in the prepress industry.

As a result of attempts to introduce more intuitive color manipulation, HSL (Hue, Saturation, Luminance) changes are executed in three steps. In the first step, the original image is translated from the printable CMY color space to the HSL (or equivalent) color space, which is unprintable, but which is convenient for manipulation. In the second step, the HSL changes are applied according to a binary mask and are executed in the new color space. In the third step, the new image is translated back to the printable color space. The original black remains unchanged in such color manipulation.

In this approach it is difficult to: 1) accomplish quality translations between color spaces; 2) meet precision requirements; 3) decide how to treat unprintable colors; 4) achieve interactive performance; 5) avoid the undesirable effects of binary weighting; and 6) not distort the original gray balance.

It is an object of the invention to solve the above identified problems, and particularly not to distort the original gray balance.

According to the method of the invention, a new value of K (black) is defined corresponding with a change in CMY components. Original CMYK printing colors are provided. The original CMY printing colors are corrected to create new CMY printing colors. A value of the smallest one of the original CMY colors is sensed. A value of a smallest one of the new CMY colors is sensed. By use of each of these two smallest values, the original K printing color value is modified to create a recalculated K printing color value.

Color manipulation is simpler and more natural when done by hue, saturation and luminance adjustments. The restoration of the gray balance is intuitively expected. The hue, saturation and luminance changes are executed in one step without the color space translations. The restoration of the black balance is based on the original CMYK image combined with the CMY changes applied to it during the color manipulation.

Hue, saturation, and luminance changes are applied directly to the CMY colors and are executed smoothly in the CMY color space according to a weighted mask. After the CMY color manipulation, the new black component is created to restore the gray balance of the original image. The image remains always in the same printable CMYK color space.

All calculations are based on the original CMYK image and are executed on a pixel by pixel basis with respect to the nonlinear gray line rather than to the color cube diagonal (the gray line has no color, but represents changing luminance). The gray line extends nonlinearly through the CMY color space. Both original and new images are maintained in the printable CMY color space.

After the CMY color manipulation, the new black component is created to restore the gray balance of the original image. The image always remains in the same printable CMYK color space.

The restoration of the gray balance is based on the original CMYK image combined with the CMY changes applied to it during the color manipulation. The new black calculation is executed in one step and is performed on a pixel by pixel basis. Both original and new images remain in the printable CMYK color space.

K treatment for arbitrary CMY changes is provided. The following change, applied directly to CMY colors and executed in the CMY color space during the color manipulation, is given:

The new black component is created from the original black combined with the skeleton black gradation functions of the CMY components before and after the color manipulation.

FIG. 1 is a block diagram showing the method of the invention for defining a new value of K corresponding with a change in CMY components;

FIG. 2 is an illustration of a CMY color space showing hue change;

FIG. 3 is an illustration of a CMY color space showing saturation change; and

FIG. 4 is an illustration of a CMY color space showing luminance change.

The K recalculation method according to the invention is most easily understood as shown by the block diagram in FIG. 1. An image scanner creates original values C.sub.orig, M.sub.orig, Y.sub.orig, and K.sub.orig. The image scanner 1 creates these color values by scanning an original image. The color values are then input into a CMY transformation 2 where new color values C.sub.new, M.sub.new, and Y.sub.new are created based on HSL changes input into the CMY transformation 2. A masking function f(d) is also input. The new color values C.sub.new, N.sub.new, and Y.sub.new are then input into color computer 3 which converts these colors to RGB video signals for input to a color video display 4. Paper type can also be input into the color computer. The above was described in greater detail in my copending patent application, Hill Case No. P-92,1545, titled "HSL CORRECTIONS IN CMY COLOR SPACE" and incorporated herein by reference. According to the present invention, the black value K.sub.orig is input into a K recalculation 5. The minimum value selection 6 selects the smallest of the three values C.sub.orig, M.sub.orig, and Y.sub.orig. The minimum value selection 7 also selects the smallest of the new color values C.sub.new, M.sub.new and Y.sub.new. These two minimum values are then input into the K recalculation 5 which then recalculates the K and outputs K The HSL changes and K treatment shall now be described in greater detail.

As shown in FIG. 2, each hue change is applied directly to the CMY colors and is executed in the CMY color space 10. The original color rotates in its luminance plane 13 (:perpendicular to the color cube diagonal 14) around the intersection of this plane with the gray line 12 (the gray line represents absence of color but increasing luminance):

A maximum desired hue change is described as:

Input data consists of: 1) the image as an array of pixels, with the (CMY).sub.original color components specified for each pixel; and 2) an optional weighting mask as an array of pixels, with a percentage of the maximum desired change specified for each pixel:

For each pixel of the image, the following five steps are carried out:

1. Two change coefficients are calculated according to the maximum desired hue change and an arbitrary function of the optional weight: ##EQU1##

2. The intersection of the color cube diagonal with a luminance plane perpendicular to it is calculated as:

3. The intersection of the gray line with a luminance plane is calculated as the empirically established function of the color cube diagonal intersection with the same plane:

4. New color components (after the hue change) are calculated as:

5. The new color components can be clamped between the color limits of the available CMY color space and/or are subjected to the saturation adjustment which preserves the desired hue and luminance values for 0.ltoreq.C.sub.gray .ltoreq.C.sub.max ;

where COEFF is the smallest of: ##EQU2## where

6. The empirically established skeleton black gradation function for the smallest of the original CMY color components is calculated as:

7. The empirically established skeleton black gradation function for the smallest of the new CMY color components is calculated as:

8. The new black component is calculated as:

As shown in FIG. 3, each saturation change is applied directly to the CMY colors and is executed in the CMY color space 10. The original color moves along the line 15 which connects this color to the intersection of its luminance plane with the gray line 12. Decreases and increases in saturation are achieved by moving to and from this intersection, respectively:

A maximum desired saturation change is described as a percentage of the actual saturation: 0.ltoreq.CHANGE.ltoreq.1 for saturation decrease and 1<CHANGE for saturation increase. The input data consists of: 1) the image as an array of pixels, with the (CMY).sub.original color components specified for each pixel; and 2) an optional weighting mask as an array of pixels, with a percentage of the maximum desired change specified for each pixel: 0<d<100% (for a global change d=100%).

For each pixel of the image, the following five steps are carried out:

1. The change coefficient is calculated according to the maximum desired saturation change and an arbitrary function of the optional weight:

2. The intersection of the color cube diagonal with a luminance plane perpendicular to it is calculated as:

3. The intersection of the gray line with a luminance plane is calculated as the empirically established function of the color cube diagonal intersection with the same plane:

Y.sub.gray =f.sub.Y (D),,M.sub.gray =f.sub.M (D) and C.sub.gray =f.sub.c (D)

4. New color components (after the saturation change) are calculated as:

5. The new color components can be clamped between the color limits of the available CMY color space and/or subjected to the saturation adjustment which preserves the desired hue and luminance values for 0.ltoreq.C.sub.gray .ltoreq.C.sub.max ;

where COEFF is the smallest of: ##EQU3## where

6. The empirically established skeleton black gradation function for the smallest of the new CMY color components is calculated as:

7. The empirically established skeleton black gradation function for the smallest of the new CMY color components is calculated as:

8. The new black component is calculated as:

As shown in FIG. 4, each luminance change is applied directly to the CMY colors and is executed in the CMY color space 10. The original color moves along the line 16 which connects this color to the point of minimum luminance on the gray line 12. Decreases and increases in luminance are achieved by moving to and from this point, respectively:

A maximum desired luminance change is described as a percentage of the actual luminance: 0.ltoreq.CHANGE.ltoreq.1 for luminance decrease and 1.ltoreq.CHANGE for luminance increase.

The input data consists of: 1) the image as an array of pixels, with the (CMY).sub.original color components specified for each pixel; and 2) an optional weighting mask as an array of pixels, with a percentage of the maximum desired change specified for each pixel: 0.ltoreq.d.ltoreq.100% (for a global change d=100%).

For each pixel of the image, the following five steps are carried out:

1. The change coefficient is calculated according to the maximum desired luminance change and an arbitrary function of the optional weight:

2. The intersection of the color cube diagonal with the largest color luminance plane is calculated as:

3. The point of minimum luminance on the gray line is described as:

4. New color components (after the luminance change) are calculated as:

5. The new color components can be clamped between the color limits of the available CMY color space and/or is subjected to the saturation adjustment which preserves the desired hue and luminance values for 0.ltoreq.C.sub.gray .ltoreq.C.sub.max ;

where COEFF is the smallest of: ##EQU4## where:

The intersection of the color cube diagonal with the new color luminance plane is calculated as:

The intersection of the gray line with the new color luminance plane is calculated as the empirically established function of the color cube diagonal intersection with the same plane:

6. The empirically established skeleton black gradation function for the smallest of the original CMY color components is calculated as:

7. The empirically established skeleton black gradation function for the smallest of the new CMY color components is calculated as:

8. The new black component is calculated as:

For each of the above hue, saturation and luminance changes, as shown in FIG. 1, the new black component is created from the original black combined with the skeleton black gradation functions of the CMY components before and after the color manipulation:

Summarizing, the input data consists of the image as an array of pixels, with the (CMYK).sub.original color components specified for each pixel. For each pixel of the image, the following three steps are carried out:

1. The empirically established skeleton black gradation function for the smallest of the original CMY color components is calculated as:

2. The empirically established skeleton black gradation function for the smallest of the new CMY color components is calculated as:

3. The new black component is calculated as:

This new black value is clamped between .phi. and some max value.

Although various minor changes and modifications might be proposed by those skilled in the art, it will be understood that we wish to include within the claims of the patent warranted hereon all such changes and modifications as reasonably come within our contribution to the art.