Computer graphics processing and selective visual display system – Computer graphics processing – Attributes
Reexamination Certificate
1998-11-03
2001-12-04
Luu, Matthew (Department: 2672)
Computer graphics processing and selective visual display system
Computer graphics processing
Attributes
C345S593000
Reexamination Certificate
active
06326977
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to image processing and more particularly, to the construction of a perceptually optimized color palette for use in displaying an original image of higher color resolution on a device with lower color resolution.
Color images are commonly processed by computers and other digital electronic systems as bitmaps. A bitmap is an array of data representing the color of each picture element or pixel of the image. Since data is required to represent each pixel of an image, the memory necessary to store the image and the time necessary to process the image are determined both by the total number of pixels making up the image and the number of data bits used to describe the color of each pixel, known as the bit depth. The color resolution of a system or subsystem is the total number of colors which can be simultaneously used to reproduce the image. The color resolution is a function of the bit depth. The number of colors that can be simultaneously displayed (C) is given by the function C=2
N
, where N is the bit depth. For a system with a bit depth of 24, the color of a pixel can be selected from 16.7 million colors while only 256 colors are available for each pixel in a system with a bit depth of eight bits. Image quality improves with greater bit depth since the perceived “true color” of an image pixel is more likely to be found in a palette of 16.7 million colors than in a palette of 256 colors. On the other hand, three times more memory is required to store an image having 24 bit resolution than eight bit resolution. Likewise, greater bit depth increases the processing requirements and may not be useful with some display technologies, such as LCD devices. The ability to reproduce images at lower color resolutions than the color resolution of the original image is important to reduce system costs, especially for portable LCD-based devices.
Reproducing an image at a lower color resolution requires construction of a palette of colors from which the system can select a color to display as a facsimile of the original pixel color. In other words, in a system with eight bit color only 256 colors will be available to replicate the color of each pixel of an image even though the original color of the pixel may have been selected from a palette of more than 16.7 million colors. A principal objective in constructing a palette is to optimize the selection of colors so that images reproduced at a lower resolution are perceived to contain the same colors or colors that are very close to those of the original image. Several methods of optimizing the selection of colors for a palette have been developed, as described below.
One approach to selecting the colors for a palette is to determine which colors appear most frequently in an image and to fill the palette with the most frequently appearing colors. For example, in an 8 bit system the 256 colors appearing most frequently in the image would be used in the palette. The color of each pixel is mapped to the closest color in the palette when the image is reproduced. This approach may produce a high quality reproduction because any single image is likely to contain relatively few colors and it is likely that the true color of the pixel, or a color very near the pixel's true color, will be in the palette. However, image dependent palettes are usually undesirable when several images are presented simultaneously, as in a windowing environment, or when images are presented in rapid succession because palette switching is visually disruptive and computationally intensive. Braudaway, U.S. Pat. No. 4,907,075, and White et al., U.S. Pat. No. 5,130,701, both describe such systems and are incorporated by reference herein.
Another approach is to construct an image independent palette. One method of constructing an image independent palette is to truncate the high bit color representation of the pixel's color. In this approach the least significant digits of the triplets making up the high bit color representation are discarded until the number of bits remaining corresponds to the system's bit depth. A substantial part of the information provided by the original high bit color description is lost in the truncation process. In addition, the color produced by the display can vary substantially from the pixel's true color.
A complicating factor in choosing optimal colors for a palette for a lower color resolution image reproduction is that more than one system is used to describe color in digital systems. Generally, output devices for digital systems require that a color be defined in terms of the intensities of its red, green, and blue (RGB) components. The RGB color model is a cube in a Cartesian coordinate system with red usually being on the x-axis, blue usually being on the z-axis, and green usually being on the y-axis. Black (minimum intensity of red, green, and blue) is located at the intersection of the red, green, and blue axes and white (maximum intensity of red, green, and blue) is located at the corner diagonally up and across the RGB cube from black. The diagonal between the black and white corners of the RGB color gamut is the gray scale. Any color in the cubic gamut of colors is defined by a point located by the red, green, and blue coordinates. RGB color description is suitable for monitors because the cathode ray tube (CRT) composes each pixel from three subpixels; one having a principally red phosphor, one having a principally green phosphor, and one having a principally a blue phosphor. Alternatively, some monitors include a “white” phosphor together with suitable filters to provide RGB output. In a system with 24 bit color depth the color descriptor is composed of an ordered triplet specifying the intensities of red (8 bits), green (8 bits), and blue (8 bits). The intensity of each of the colors (red, green, and blue) is mapped to a memory which outputs the correct voltage to each of the three electron guns in the CRT causing the three phosphors to output the correct light intensity to achieve the specified color for the pixel. Unfortunately, the RGB mapping is device dependant because the actual color produced depends upon the device. To be optimized for human perception, the color palette must take into account the nonlinear relationship of input and output intensity of the display device which is generally referred to as the gamma function of the device.
A color model utilized in many digital video systems is the YC
B
C
R
(YUV) color model. The YC
B
C
R
color space also describes color using an ordered triplet. The ordered triplet specifies three different parameters, namely, luminance (Y), chrominance-blue (C
B
), and chrominance-red (C
R
). The YC
B
C
R
color model is also a cube at the intersection of three axes of a Cartesian coordinate system. As with the RGB color model, black (minimum luminance, chrominance-blue, and chrominance-red) in the YC
B
C
R
color gamut cube is located at the intersection of the axes of the color space. In the YC
B
C
R
color model white (maximum luminance, minimum chrominance blue and minimum chrominance red) lies on the Y-axis. In recent years, the YC
B
C
R
color model has become common for image description because research has shown that significant improvement in the compressibility of image data is possible with this color representation. Image compression reduces the memory requirements to store the image and the bandwidth requirements to transmit it. JPEG, MPEG, and H.261/H.263 image and video compression standards all use YC
B
C
R
color components. Digital television also uses YC
B
C
R
color components.
YC
B
C
R
color data can be converted to RGB data, and vice versa, by a matrix transformation. If there is sufficient bit depth a simple mathematical transformation may be satisfactory to transform between YC
B
C
R
and RGB. However, when a color palette is used the matrix transformation does not produce an optimal palette. When the RGB color gamut is superimposed on the YC
B
C
R
color gamut, the RGB and YC
B
C
Chernoff Vilhauer McClung & Stenzel LLP
Harrison Chante′
Luu Matthew
Sharp Laboratories of America Inc.
LandOfFree
Rendering of YCBCR images on an RGS display device does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Rendering of YCBCR images on an RGS display device, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Rendering of YCBCR images on an RGS display device will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2574359