Image analysis – Color image processing
Reexamination Certificate
2000-12-06
2004-07-20
Tran, Phuoc (Department: 2621)
Image analysis
Color image processing
C382S276000
Reexamination Certificate
active
06766051
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to the art of color image processing. For example, it finds application where an image created on, or prepared for rendering on, a first or source device is prepared for rendering on a second or destination device.
2. Description of Related Art
When an image is prepared for rendering on an electronic device the image is represented as a set of pixels. Each pixel describes a small portion of the image in terms of colorant pixel values for the colorants available on the rendering device. For example, typically a cathode ray tube (CRT) based computer display screen is comprised of red (R), green (G) and blue (B) phosphors. An image prepared for display on a CRT is described with a set of pixels. Each pixel describes the intensity with which the red, green and blue phosphors are to be illuminated on a small portion of the CRT. A similar procedure is followed when an image is prepared for rendering on a printing device. Currently, at least some color printing devices apply cyan (C), magenta (M), yellow (Y), and sometimes black (K) colorants to a print medium, such as paper or velum, in order to render an image. Such printing devices are said to operate in a CMY or CMYK color space. When an image is prepared for rendering on a color-printing device, the image is represented as a set of pixels. Each pixel describes a small portion of the image by calling for an appropriate mixture of the available colorants. Typically, the pixel value for each colorant can range from 0 to 255. The higher a colorant's pixel value is, the more of that colorant the color image processor applies to the print medium. In a system employing 8-bit precision for the colorant signals, the number 255 represents the maximum or fully saturated amount of colorant. The number 0 is used when none of a particular colorant is required.
In a CRT operating in RGB (red, green blue) space, fully saturated red is described by a pixel calling for R=255, G=0, B=0. In a printer operating in CMYK (cyan, magenta, yellow, black) space, fully saturated red is described by a pixel calling for C=0, M=255, Y=255, K=0 Magenta and yellow colorants combine through simple subtractive mixing and are perceived as red. There is no guarantee that the red described in RGB space and displayed on the CRT is the same red described in CMYK space and printed on a page. In fact, it is quite likely that the spectral properties of the red phosphor used in the CRT will be different than the spectral properties of the subtractively mixed magenta and yellow colorants of a particular printer.
As mentioned above, the CRT and the CMYK printer use different materials to generate the perception of color. The materials used impact the set of colors that each device can reproduce.
The set of colors a device can produce is referred to as the color gamut of the device. There is no guarantee that a color that can be produced by a first device is within the color gamut of a second device. This is even true when both devices are CMYK printers.
Where color matching is required between two devices such as the CRT operating in RGB space and the printer operating in CMYK space, transforms based on calibration and measurement are required. In such a situation it is possible, for example, that the pure red RGB CRT pixel mentioned above, is mapped to a CMYK printer pixel calling for a less than fully saturated magenta component and a small amount of a cyan component. For example, the CMYK version of the original RGB red pixel referred to above might call for C=27, M=247, Y=255, K=0. Furthermore, if one wants to print a copy of the original pure red RGB CRT pixel on a second printer it is quite likely that a second transform will have to be used. That transform may translate the original RGB CRT pixel to a second CMYK pixel. For example, the second transform may map the original RGB CRT pixel to a second CMYK pixel calling for C=20, M=234, Y=240, K=35. One reason two different CMYK printers may require different transforms is that different printers use different colorants. For example, a first magenta colorant used in a first printer may have a different spectral content than a second magenta colorant used in a second printer. Likewise, a first yellow colorant used in a first printer may have a different spectral content than a second yellow colorant used in a second printer.
From the foregoing discussion it can be seen that an image prepared for rendering on a first device may need to be transformed if it is to be properly rendered on a second device. Such a transformation is an attempt to emulate the first or source device onto the second or destination device. In order to achieve calorimetric matching, the emulation of the color gamut of the CRT on the first CMYK printer caused the red CRT pixel to be mapped to a first CMYK pixel calling for C=27, M=247, Y=255, K=0. The emulation of the color gamut of the CRT on the second CMYK printer caused the red CRT pixel to be mapped to the second CMYK pixel calling for C=20, M=234, Y=240, K=35. Obviously, therefore, even where there is no RGB CRT image involved, an image prepared for printing on the first printer may have to be transformed before its calorimetric content can be matched on the second printer. In such a situation the first printer is said to be emulated on the second printer.
For example, when, a photographic image has been prepared for rendering on a first CMYK device, for example a Standard Web Offset Printing (SWOP) device, but must then be rendered on a second CMYK device, for example, a xerographic printer, a “4 to 4” transform is typically used to emulate the first device on the second device.
An image processor must be able to satisfactorily transform images, prepared for rendering on a staggeringly large number of source devices, for rendering on an equally large number of rendering devices. For example, a personal computer, configured to operate as an image processor, through the use of, for example, a web browser, word processor, or desktop publishing software, must be able to accept images created or downloaded from almost anywhere, and render the images pleasingly on any image output terminal the user connects to the personal computer. For example, the image processor must be able to pleasingly display images on a computer screen and send commands to have images rendered pleasingly on any printer the user connects to the personal computer. Similarly, document processors and copiers must transform images pleasingly and accurately. Indeed, in document processors and copiers, color accuracy requirements can be very stringent. For example, in a large print job, comprising a corporate sales presentation, it is very important that colors in a company logo be reproduced accurately and consistently. Similarly, colors in a clothing catalog must match the colors on depicted garments.
Typically transformations are carried using look up tables. For example, an image prepared for rendering in a RGB color space of a computer monitor, is transformed into a CMYK color space of a printer, using a three-dimensional look up table. Each of the RGB parameters is used as an axis or index into the look up table. The RGB parameters are typically represented by 8-bit values for each of the R, G, and B components. Such a 24-bit RGB vector is capable of addressing over 16.77 million locations or describing over 16.77 million colors. However, look up tables of this size are physically impractical in most cases. Therefore, such look up tables are typically partitioned into a smaller size, such as, for example, 16×16×16 (4096) table locations or nodes. Each node stores, for example, a CMYK value. CMYK values of points not directly represented by nodes, are determined by a form of interpolation among nodes of the look up table (LUT).
An illustration of the operation of a prior art
Balasubramanian Thyagarajan
Klassen R. Victor
LandOfFree
Adaptive tree-base lookup for non-separably divided color... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Adaptive tree-base lookup for non-separably divided color..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Adaptive tree-base lookup for non-separably divided color... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3187639