Image analysis – Color image processing – Color correction
Reexamination Certificate
2000-01-19
2003-10-21
Wu, Jingge (Department: 2623)
Image analysis
Color image processing
Color correction
C358S518000, C345S604000
Reexamination Certificate
active
06636628
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of Invention
This invention relates to color reproduction.
2. Description of Related Art
Generally, colors are represented in two ways, in device dependent color spaces and in device independent color spaces. To illustrate, most color display monitors, such as, for example, color computer monitors, display colors in the red/green/blue (RGB) color space, i.e., with respect to the amount of red, green, and blue that a particular displayed color contains. Using this technique, the color yellow, for example, is displayed on a color display monitor by combining a red image value of 100 percent red with a green image value of 100 percent green and a blue image value of zero percent.
Furthermore, the red, green, and blue (RGB) color values associated with the particular colors for a color display monitor are device dependent. This means that the RGB values associated with a particular color, viewed on a specific color display monitor, are unique to that specific color display monitor or, at least, to that brand of color display monitor. Simply put, because RGB color values are device dependent, if identical RGB color values, such as, for example, a red image value of 100 percent red, a green image value of 100 percent green, and a blue image value of zero percent, are input and displayed on two different color display monitors, the resulting yellow color displayed on the two color display monitors will probably not appear exactly alike.
Similarly, most color marking devices, such as, for example, color printers, print colors in device dependent terms. However, unlike most color display monitors, most color marking devices use a cyan, magenta, yellow, and black (CMYK) color space, i.e., a combination of cyan (C), magenta (M), yellow (Y) and black (K) (CMYK) to arrive at the color marking device's printed colors. Consequently, as with RGB color values, CMYK color values are device dependent. Thus, as described above with respect to colors being displayed on color display monitors, if identical CMYK colors are printed by two different color marking devices, the printed colors will probably not appear exactly alike.
The other way of describing color is in device independent color spaces. By describing color in a device independent color space, consistent colors can be reproduced regardless of the type of device that is used to display or print the color. Therefore, color reproduction is generally done by defining colors in a device independent color space, such as, for example, L*a*b*, X Y Z, or L h v.
In an attempt to provide accurate color matching between color display devices and color marking devices, various color matching techniques have been developed that use models to translate colors from one color space to another color space. These models usually manifest themselves in the form of predetermined multi-dimensional look-up tables. These predetermined multi-dimensional look-up tables, such as, for example, a look-up table with inputs and outputs containing more than one dimension, translate colors from one color space to another color space while attempting to maintain the translated color's perceived appearance. For example, if a user creates an image on a color display monitor and subsequently prints the created image without any color matching, the colors observed on the printed image may differ significantly from the colors originally observed on the color display monitor. However, if some type of color matching model is used, the discrepancies between the colors originally observed on the color display monitor and the colors observed on the printed image can be reduced.
Generally, to solve the problem of color matching, a printer inverse is produced for controlling colors in device dependent or device independent color spaces. Colors in the printer color space are usually in a device independent color space. The printer inverse is required to interpolate irregularly sampled multidimensional color data and is normally obtained for a particular color marking device by performing experimentation on the particular color marking device. The printer inverse is a multidimensional look-up table, preferably structured in such a way that the input nodes are regularly spaced, with the nodes located on a sequential plane.
In particular, the printer inverse is a look-up table that converts colors from device independent color spaces to device dependent color spaces, such as, for example, from L*a*b* to CMYK, or from XYZ to CMYK. For networked printers that utilize a print driver that operates in an unascertained device independent color space, the printer inverse can operate as a look-up table to convert output L*a*b* color space to input L*a*b* color space. The look-up tables are generated by measuring the printer forward transfer function between the inputs into the printer and the outputs from the printer.
For example, in a PostScript® print path, the PostScript® interpreter with colors in the XYZ/L*a*b* device independent color space becomes the input and the corresponding colors measured on a resulting color print becomes the output. The colors of the resulting color print can be measured by a color sensor, such as, for example, a spectrophotometer. The forward transfer function is then used to create the printer inverse and the associated rendering intents, such as, for example, calorimetric, pictorial/perceptual, saturation, pure, and the like.
SUMMARY OF THE INVENTION
Although the printer inverse can be generated by simply swapping the data of the forward transfer function, the data of the input grids for a printer inverse generated by merely swapping the data becomes unstructured because the output grids of the forward transfer function are unstructured. When a swapping type of printer inverse is utilized, because of the unstructured nature of the input/output (I/O) values in the swapped printer inverse look-up table, an efficient method for obtaining an effective mapping table, or printer inverse, must be used so that the relationship between the input nodes and the output nodes of the printer inverse look-up table is structured. Multidimensional interpolation is a key to obtaining such a structured mapping table.
Many techniques have been proposed for interpolating multidimensional, unstructured look-up tables. For example, Shepard's interpolation as disclosed in Donald Shepard, “A Two Dimensional Interpolation Function for Irregularly Spaced Data”, ACM National Conference Proceedings, Page 517-524, 1968. Additionally, the moving matrix method as disclosed in Raja Balasubramanian, “Refinement of Printer Transformations Using Weightless Regression”, Xerox Digital Imaging Technology Center, Webster, N.Y. 14580. Furthermore, master color controls (MCC) can be used to interpolate multidimensional, unstructured look-up tables, as disclosed in U.S. patent application Ser. No. 09/083,203, incorporated herein by reference in its entirety.
The incorporated
203
application discloses a method of reducing and controlling color drift between a desired image, and an output image printed by a marking device that is intended to match the desired image, by detecting a current output color in the output image with a color sensing device. A difference between the current output color in the output image and a corresponding color in the desired image is then determined. A next output color in the output image is then automatically set equal to a corrected color that minimizes the difference between the next output color and the corresponding color in the image. This is preferably done on a real-time basis.
However, Shepard's algorithm is time consuming and is not accurate when compared to other methods. Although computational time is not an issue in the moving matrix method, the range and the interpolation accuracy is relatively poor when compared to a master color controls (MCC) method. The master color controls (MCC) method uses an algorithm incorporating a purely control based technique with a multi-input/multi-output controller. The
Dianat Sohail A.
Mestha Lingappa K.
Viassolo Daniel Edgardo
Wang Yao Rong
Wu Jingge
Xerox Corporation
LandOfFree
Iteratively clustered interpolation for geometrical... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Iteratively clustered interpolation for geometrical..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Iteratively clustered interpolation for geometrical... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3112434