Facsimile and static presentation processing – Static presentation processing – Attribute control
Reexamination Certificate
1999-10-14
2001-12-18
Williams, Kimberly A. (Department: 2622)
Facsimile and static presentation processing
Static presentation processing
Attribute control
C358S528000, C358S525000
Reexamination Certificate
active
06331902
ABSTRACT:
BACKGROUND
The present invention relates to digital image processing including scaling and color conversion of RGB signals to CMYK format, such as for printing high quality images from digital cameras and the like.
In the processing of graphic images, it is often desired or required to alter the magnification. Traditionally, optical means such as interchangeable lenses, zoom lenses, and mechanical positioners have been used on copy cameras and photo-copiers of the prior art. More recently, photo-copiers have incorporated linear sensor arrays that are sequentially and repetitively read in a main scan direction as the arrays are moved across the page in a sub-scan direction. Numerous systems have been developed for selectively changing the magnification electronically, such as by using a computer programmed with appropriate software. Such systems are prohibitively slow in many applications. U.S. Pat. No. 4,920,571 to Abe et al. discloses an image processor that incorporates an interpolator that uses look-up tables for adding image data in enlarging operations and selectively reducing the data in reducing operations. In one example, the magnification can range by {fraction (1/64)} ths from {fraction (32/64)} to {fraction (128/64)} (0.5-2.0). The processor of Abe et al. has the disadvantage of requiring a separate look-up table for each magnification. Thus it is ineffective in that it is capable of only a limited choice of magnifications; and/or it is overly complex, requiring a multiplicity of look-up tables.
U.S. Pat. No. 5,535,007 to Kim discloses a pipeline processor for enlarging or reducing an original image to obtain a copy image scaled in 1 percent increments, with sharpening and color conversion. The device includes means for inserting or removing pixel data in a main-scanning direction and inserting or omitting the scanning lines using separate look-up tables that are populated according to designated image input parameters and control factors that correspond to a desired magnification, which can typically range from 50% to 400%. Also disclosed are conversion from RGB to CMYK color formats by a CMY look-up table that feeds an under color removal (UCR) comparator and subtractor, and a color corrector that selectively substitutes converted color reference values for C, M, and Y, depending on a comparison of input data with respective register values. The processor of Kim, while enabling scaling in one-percent increments, has the disadvantage of having to populate the look-up tables for each different magnification, thereby adding complexity to the processor and slowing the process.
U.S. Pat. No. 5,867,634 to Hirota et al. discloses an image processor in which an image of a document is read by an image sensor, and a reduction ratio of the image is set. Matrix interpolation (for reducing magnifications) is then performed on the image data between a pixel under interest and pixels adjacent thereto as functions of the reduction ratio. Next, reduction/magnification is carried out by sequentially reading stored interpolation data at a read clock rate that is proportional to a write clock rate at which the interpolation data is stored. The data is read into eight smoothing filters that are configured for separate integer ranges of positive magnification, the data from one of the filters being selected as output depending on the actual magnification. Color conversion is done in stages, first in logarithmic conversion by table look-ups from reflective to density RGB values, followed by UCR/BP processing and color correction by matrix masking. The image processor of Hirota, unfortunately, is excessively complex, requiring a number of variable multiplication operations (including squaring) to be performed on multiple channels of data for each pixel to be processed. Also, the scaling subsystem effectively requires images to be greatly enlarged, then reduced.
Thus there is a need for an image processing system that overcomes the disadvantages of the prior art.
SUMMARY
The present invention meets this need by providing an image processing system having variable magnification by adapted bi-linear interpolation at one of a predefined set of scale factors, and dropping an equally spaced subset of pixels in cases of other scale factors. Optionally or alternatively, the system provides color conversion from RGB to CMYK with correction/matching using rectilinear interpolation. In one aspect of the invention, a system for scaling an input image defined by an array of source pixels to produce an array of scaled pixels includes a memory for storing the source pixels; means for receiving a scale factor within a range of specifiable scale factors; means for selecting an interim scale factor from a numerically spaced set of predefined main scale factors, the main scale factors being spaced from a low end of the range, one of the scale factors being at a high end of the range, the interim scale factor being that nearest the received scale factor in the direction of the high end of the range; means for determining, for the interim scale factor, a pixel transformation between a subset of the source pixels and a corresponding subset of intermediate pixel data as scale functions of source pixel values; means for producing the intermediate pixel data from the source pixels using the interim scale factor and corresponding scale functions; and means for interpolating the intermediate pixel data to produce the scaled pixels by dropping an evenly spaced subset of the intermediate pixel data when the received scale factor is less than the interim scale factor.
The range can be from not more than 10 percent to not less than 200 percent, there being at least 5 of the predefined scale factors. The received scale factor can be specifiable in 1 percent increments. The predefined scale factors can be spaced by not more than 25 percent for scale factors not exceeding 150 percent. The range can extend to 400 percent, the predefined scale factors including percentages of 25, 50, 75, 100, 125, 150, 200, 250, 300, and 400. The image can be a color image, each of the pixels having at least three color components. The color components can include red, green, and blue.
The subsets of source pixels can be square arrays having a source dimension selected from the set consisting of 1, 2, and 4, the subsets of intermediate pixel data being also square arrays having a destination dimension selected from the set consisting of 1, 2, 3, 4, and 5, the destination and source dimensions having a ratio matching the predefined scale factors. The destination and source dimensions can define the ratio as a reduced fraction. The functions can be of support subsets of the source pixels, each support subset including a corresponding source subset and having a support dimension being not more than one greater than the source dimension.
The system can further include means for enhancing the scaled image using a weighting factor applied to values of diagonally adjacent pixels of respective center pixels. The means for enhancing can include means for augmenting an initial value of the center pixel by an amount proportional to the weighting factor and diminishing the value of the center pixel by an amount proportional to an average value of the diagonally adjacent pixels. The weighting factor can be variable in response to external input.
In another aspect of the invention, a system for converting a color image defined by an array of source pixels having red, green, and blue color components to produce an array of output pixels, each of the output pixels having cyan, magenta, yellow and black components, includes: the memory for storing the source pixels; a 3-D lookup table containing CMYK values corresponding to respective RGB values; means for addressing the table with high-order value portions of the enhanced pixels to obtain corresponding sets of intermediate CMYK values, the addresses corresponding to high-order value portions of the RGB values, selectively incremented or unchanged to obtain neighboring high-order value portions,; and mean
Match Lab, Inc.
Sheldon & Mak
Williams Kimberly A.
LandOfFree
System and method for digital color image processing does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for digital color image processing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for digital color image processing will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2597158