Image analysis – Image transformation or preprocessing – Changing the image coordinates
Reexamination Certificate
1996-08-02
2002-07-02
Boudreau, Leo (Department: 2621)
Image analysis
Image transformation or preprocessing
Changing the image coordinates
C382S167000, C382S251000, C382S274000, C358S451000, C358S518000, C358S520000, C358S525000
Reexamination Certificate
active
06415065
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an image processing apparatus effecting color processing with tables, and a method therefor.
2. Related Background Art
There have been proposed various color conversion technologies, in the output of color information entered for example from an image input device to a color printer or the like, for converting such color information from a device-in-dependent color space into device-dependent color components specific to the color printer. The color components entered from the image input device are three components of red (R), green (G) and blue (B), while the output color components are cyan (C), magenta (M) and yellow (Y) specific to the coloring materials in the printer. If the black color cannot be represented with these three colors, the representation is made with four colors including black (K). The conventional signal flow from the RGB input signals to the CMYK output signals is shown in FIG.
25
. The input color component signals may be RGB signals based on the NTSC or PAL standards, or can also be L*a*b* signals based on the uniform color space. Means
201
for logarithmic conversion and gamma correction prepared CMY complementary color components by logarithmic conversion. (The CMY components prepared in this step are represented by C
0
, M
0
, Y
0
respectively.) Then means
202
effects undercolor removal (UCR) and black generation to obtain a black (K) component. Then masking means
203
effects conversion to a device-dependent color space, matching the coloring materials specific to the printer. For this conversion, there has been proposed a method of calculating the conversion coefficients, utilizing a black box model. The equations for this conversion can be, for example, linear ones utilizing a 3×3 conversion matrix or non-linear ones involving higher-order terms for improving the precision.
There is also conceived a configuration including a K term in the input of the masking means, or a configuration including the UCR itself in the masking means. Output gamma correction means
204
corrects the four color components according to the characteristics of the printer. Pseudo gradation process means
205
effects a pseudo gradation process according to the number of levels generatable by the printer, and the obtained image data are transmitted to a printer engine and printed therein.
In the conventional configuration shown in
FIG. 25
, there has been explained conversion means employing linear or non-linear approximation, but, in recent years, there is being principally employed methods utilizing three-dimensional color correction table for more precise conversion. For example the Japanese Patent Laid-Open Application No. 63-2669 proposes a color conversion method by so-called direct mapping, utilizing a color correction table based on all the combinations. Also there have long been proposed various methods of preparing a table with quantized lattice points of a limited number, instead of preparing all the combinations, and effecting color conversion for the input value other than the lattice points by an interpolating calculation. As an simplest example, there will be explained interpolation by 8 vertices of a cube, with reference to
FIG. 26
, showing the method of interpolation for an input point i, based on vertices a-h of a cube, contained in the lattice points stored in the table. The table contains the conversion information f for these vertices (f(a) to f(h) for the points a-h), the value g(e) of the point e after conversion is given by:
g(e)=(1−x)(1−y)(1−z)f(a)+x(1−y)(1−z)f(b)+(1−x)(1−y)zf(c)+x(1−y)zf(d)+(1−x) y(1−z)f(e)+xy(1−z)f(f)+(1−x)yzf(g)+xyzf(h)
This method can achieve desired color conversion in easy manner with a limited table capacity, by linear approximation with lattice points constituting a cube.
Also as another conventional example, the Japanese Patent Laid-Open Application No. 7-30772 proposes a pseudo color conversion without interpolation, by a pseudo gradation process. This method utilizes two levels of pseudo gradation process, wherein a gradation level pre-conversion effects pseudo gradation process for converting into the coordinates of optimum lattice points, in order to eliminate the input values other than the lattice points of color conversion, and a gradation level post-conversion effects quantization matching the printer, again by a pseudo gradation process. This method is based on a concept that the image quality is not affected by the limitation of the number of gradation levels by the pseudo gradation process in the pre-conversion, if a rough quantization (for example binarization) is conducted in the post conversion.
However the conventional methods explained above are associated with the following drawbacks. The color conversion process is often executed in a printer driver software on the host computer, in case of the color ink jet printer, the melting thermal transfer printer or the sublimation thermal transfer printer. With the recent improvement in the resolution of the printer engine, the number of pixels to be processed has significantly increased, so that the printer driver software requires a long process time. For this reason it is strongly desired to reduce the process time as far as possible, without deteriorating the precision of the color conversion process. In the aforementioned conventional method of calculating the corrosion value between the lattice points by interpolation, the calculation of a point g(e) requires 24 multiplications and 7 additions, leading to an enormous process time.
Also the method of the Japanese Patent Laid-Open Application No. 7-30772, employing the gradation level pre-conversion for eliminating the input values other than the lattice points prior to the input of the lattice points by the pseudo gradation process, is faster than the interpolation mentioned above, but, unless a large number of quantization levels is selected after the pre-conversion, generates a pseudo contour in the stage of pre-conversion, so that the image quality remains deteriorated also in the post-conversion. A large number of quantization levels after the pre-conversion leads to a larger table capacity, thereby increasing the load on the memory in the host computer or a longer time for the address search for such large-capacity table.
Also for a printer engine of a higher resolution, there can be conceived, for reducing the number of pixels to be processed, a method of reducing the input resolution and effecting image expansion after the color conversion, or a method of effecting an expansion at the binarization for conversion into the output resolution. In the above-explained configuration, the color conversion has to be more precise, as the resolution at the color conversion becomes lower. The image quality and the speed or the table capacity are mutually contracting requirements, and there has not been known a color conversion process satisfying all these conditions.
The foregoing conventional methods are also associated with the following drawbacks. As an example, there is considered a system in which the input resolution is different from the output resolution. The output resolution of the printers is increasing year after year, but, if the input resolution is matched with such increasing output resolution, there will result various difficulties such as the load for the preparation and processing of the image information on the host computer, the load for the color conversion and the pseudo gradation process in the printer driver, and the load of transfer time of the image information from the printer driver to the printer. It is therefore conceived a configuration of entering the information with a low resolution, thereby alleviating the load of image processing, and preparing and releasing the image information of a high resolution matching the printer engine.
The color conversion is often execu
Boudreau Leo
Chawan Sheela
LandOfFree
Image processing apparatus and method therefor does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image processing apparatus and method therefor, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing apparatus and method therefor will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2869501