Facsimile and static presentation processing – Static presentation processing – Attribute control
Reexamination Certificate
1998-07-08
2002-11-05
Rogers, Scott (Department: 2624)
Facsimile and static presentation processing
Static presentation processing
Attribute control
C382S252000, C345S616000
Reexamination Certificate
active
06476934
ABSTRACT:
FIELD OF INVENTION
The present invention relates generally to digital image processing and, in particular, to methods and apparatus for the digital halftoning of continuous tone images.
BACKGROUND ART
Digital halftoning is a technique used in image processing to produce a halftone output image from a continuous tone input image. In order to halftone a digital color image, a continuous tone color image is sampled, for example with a scanner, and the samples are then digitised and stored in a computer device. In a full color system, the digitised samples, or pixels, can consist of independent binary representation of the color components of an image. For example, in the well known red, green and blue system (RGB), utilised in most output displays, the digitised samples or pixels consist of binary representations of the red, green and blue scanned color values respectively. These binary representations typically range from 0 to 255 for each color, thereby comprising 8 bits for each primary color, or 24 bits to represent one pixel. Another color system involves cyan, magenta, yellow and black (CMYK) and is used predominantly in printer output devices.
Standard cathode ray tube (CRT) type display devices are able to display each pixel with a large number of variations of each color component of the pixel, giving rise to the desirability of storing 256 possible values for each color component of each pixel of a CRT type display. Other output devices however, such as printers and certain display devices such as ferro-electric liquid crystal displays are often designed to only be able to display a limited number of colors or intensity values for each output color, Hence, when displaying an image on such devices, it is necessary to create the sensation of multilevel colors by suitably distributing the possible output colors in the output image. By way of example, it can be assumed that each pixel of a first example display device is able to display red, green and blue color values (RGB), with each color value taking either one of a totally on or totally off state. Therefore, each color component value can display one of two colors and the total number of colors which can be displayed by such a pixel will be 2×2×2=8 colors.
Two well known methods of halftoning are error diffusion and dithering. To describe error diffusion and dithering, consider by way of example, a monochrome display system where each pixel of an array of pixels can display either black or white. Assume the input image has 256 possible levels of display or values, 0 through 255. A decision must be taken for each pixel whether to display an “off” value of 0 or “on” value of 255.
In error diffusion for this example system, an “off” value is displayed at a current pixel when the input value at the pixel plus error values transferred to the pixel is less than or equal to 127, and an “on” value is displayed when the input value plus error values transferred to the pixel are greater than or equal to 128. An error value at the current pixel is derived, being the input value at the pixel plus the error values transferred to the pixel less the chosen output value, Portions of this error value are then transferred to surrounding pixels which have not, as yet, been output, in accordance with a known distribution mask, This has the effect of spreading, or “diffusing” the error over several pixels in the output image.
FIG. 25A
illustrates a known error diffusion mark, where a current pixel
400
is being processed and the error value at the current pixel
400
is distributed to a limited set of neighbouring pixels
401
to
406
according to values provided by the mask. In this example, 2/8 of the error is distributed to the pixel
401
and 1/8 of the error distributed to the pixel
402
, and so on.
With such arrangements, halftoning by error diffusion typically proceeds by processing pixels in raster order in the manner shown in FIG.
25
B. There, a first current pixel output value is decided using an input sum for the first current pixel, in particular, a sum of the input value for the first current pixel with a weighted sum of error values of previously processed pixels. After the first current pixel output value is determined, an output sum for the first current pixel is calculated as an input sum less the current pixel output value. Note that the output sum is a weighted sum of error values of processed pixels including the first current pixel, where the error value for the first current pixel is its input value less its output value, and the weight for the first current pixel is one. The output sum is then divided into portions which are distributed to a small group of close neighbouring unprocessed pixels. That is, each portion is added to the input sum of a neighbouring unprocessed pixel. In this way, the input sum for a second current pixel is calculated as the sum of the input value for the second current pixel plus a weighted sum of error values of previously processed pixels, and the processing applied to the first current pixel can be repeated for the second current pixel.
As such, error diffusion is a neighbourhood process halftoning technique. Expressed in another way, error diffusion provides a halftone output at a pixel that depends on the halftone output of pixels in its neighbourhood. In conventional error diffusion, the error between a halftoned output value and the input value for any one pixel is typically spread amongst a neighbourhood comprising a limited number of pixels, six (6) as seen in the above example. It follows therefore that the output value of a pixel is predominantly determined by the influence of a small neighbourhood of pixels associated therewith.
It has been found that error diffusion produces-unsatisfactory results when an input image comprises video or other forms of data having motion (ie. dynamic images) or noise characteristics. When dynamic images or images having a certain associated noise therein are error diffused, the error diffusion process is subject to slight variations from one frame to the next which in turn results in pixel values begin turned “on” and “off” at a detectable and distracting rate. This problem can be variously described as “sparkling noise”, “dancing dots” and “twinkling”, as the effect is to produce an area having rapidly changing individual pixel values but having a substantially constant overall color. This problem applies equally to both monochrome and full color display output devices in which such images are displayed.
Further, regions of low and high intensity are typically handled poorly by traditional error diffusion. This problem may also occur in both still and moving images.
The halftoning process of dithering traditionally involves the creation of “dither matrix”, where the input value of a current pixel is compared with a corresponding value in the dither matrix and an output value for the current pixel is derived, For example, if the dither matrix value is less than the input value of the current pixel, the display device, such as a printer or display, produces an “on” value at the current pixel.
A dither matrix is preferably constructed having certain characteristics which improve the appearance of output halftoned images. These characteristics include arrangement of dither matrix values so that resulting “on” or “off” pixel values are as spread out as possible.
The process of dithering dynamic images, that is, dithering images forming part of an image sequence, however results in the reproduction of images have a noisy, “mottled” appearance in regions having a substantially constant color. Another disadvantage of dithering includes poor edge sharpness at the edges of graphical objects forming part of the image.
It is an object of the present invention to provide an alternative form of image processing which leads to improved output values for display.
SUMMARY OF THE INVENTION
In accordance with one aspect of the present invention there is disclosed a method of processing an image, said image comprising a plurality of pixels
Arnold Michael Richard
Ilbery Peter William Mitchell
LandOfFree
Geometrically reducing influence halftoning does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Geometrically reducing influence halftoning, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Geometrically reducing influence halftoning will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2943143