Image analysis – Image enhancement or restoration – Variable threshold – gain – or slice level
Reexamination Certificate
1998-06-08
2001-02-27
Lee, Thomas D. (Department: 2724)
Image analysis
Image enhancement or restoration
Variable threshold, gain, or slice level
C382S251000
Reexamination Certificate
active
06195468
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method of converting continuous tone images into pseudo-halftone binary images.
2. Description of Related Art
An error-distributing image conversion method has been proposed to convert a continuous tone image into a pseudo-halftone image with high quality. The continuous tone image is comprised of a plurality of pixels, each pixel having a density defined by one of plural tone levels. The plural tone levels are defined in a density range between a predetermined minimum density and a predetermined maximum density. The error-distributing image conversion method converts this multilevel density value, at each pixel, into one of two levels, ON and OFF, for example. The plurality of pixels, each thus having a density value of either ON or OFF, construct the pseudo-halftone image.
SUMMARY OF THE INVENTION
In one conceivable process of the error-distributing image conversion method, when converting the multilevel density value of one pixel into a bilevel density value, the multilevel density value may be first modified by binary conversion errors distributed from already-processed pixels that are located in a neighborhood of the subject pixel. Then, the modified multilevel density is compared with a predetermined threshold value. In accordance with the compared result, the subject pixel is turned ON or OFF. Then, a binary conversion error is calculated as a difference between the modified multilevel density and a density represented by the determined bilevel value of ON or OFF. That is, when the subject pixel is turned ON, the conversion error is calculated through subtracting the maximum density from the modified multilevel density. When the subject pixel is turned OFF, on the other hand, the conversion error is calculated as equal to the modified multilevel density. The binary conversion error will be fractionally distributed or affected onto pixels, which are located neighboring to the subject pixel and which have not yet been processed.
In order to distribute or affect the binary conversion error to the unprocessed neighboring pixels, there have been proposed various methods, such as an error diffusion method and a minimized average error method.
The error diffusion method has been proposed by Floyd, et al. in “An Adaptive Algorithm for Spatial Greyscale” SID 17 [1976]. According to the error diffusion method, after the subject pixel is converted into a binary value, the generated binary conversion error is fractionally distributed or diffused to neighboring pixels not yet processed. Each fraction of the binary conversion error is accumulated into a conversion error sum to be used for processing a corresponding unprocessed neighboring pixel.
The minimized average error method has been proposed by J. F. Jarvis, C. N. Judice, and W. H. Ninke, in “A Survey of Techniques for the Display of Continuous Tone Pictures on Bilevel Displays”, Computer Graphics and Image Processing.5,13-40(1976). According to the minimized average error method, before processing one pixel, the multilevel density of that pixel is added with a sum of fractions of binary conversion errors, which have been generated at already-processed pixels neighboring to the subject pixel. The thus modified multilevel density is converted into the binary value.
The above-described error-distributing image conversion processes can convert a continuous tone image, whose pixels have multilevel densities, into a pseudo-halftone image, whose pixels have bilevel densities. The pseudo-halftone image can be reproduced by a bilevel printer, which is designed to selectively apply an ink dot or not at each pixel location.
The error-distributing image conversion processes, however, induce the problems described below.
It is now assumed that the conceivable error-distributing image conversion process is performed on a pixel by pixel basis onto a continuous tone image D shown in FIG.
1
. All the pixels in the continuous tone image D are processed one by one from left to right along each pixel line, i.e., in a main scanning direction x. The pixel lines are processed from top to bottom along the auxiliary scanning direction y. As the pixels are thus proceeded, binary conversion errors are obtained and distributed to unprocessed neighboring pixels. More specifically, each pixel density value is modified by fractions of errors distributed from already-processed neighboring pixels. The modified density is then compared with a predetermined threshold value. Based on the compared result, the pixel density value is converted into a binary value. A binary conversion error generated through the binary conversion process will be affected to neighboring unprocessed pixels.
As shown in
FIG. 1
, the continuous tone image D has three regions R
1
, R
2
, and R
3
which are arranged in this order along the auxiliary scanning direction y. Pixels in regions R
1
, R
2
, and R
3
are subjected to the error-distributing image conversion process in this order.
It is also assumed that density of each pixel in the continuous tone image D is defined in a multilevel density range between zero (0) and 255. The region R
1
is a non-uniform density region. Pixels with two or more different densities are distributed non-uniformly in region R
1
. Contrarily, regions R
2
and R
3
are uniform density regions. All the pixels in region R
2
have the minimum density in the multilevel density range. That is, all the pixels in region R
2
have the same density of zero (0). All the pixels in region R
3
have the same density of another value greater than zero (0). In this example, all the pixels in region R
3
have the same density of one (1).
When this continuous tone image D is converted into a pseudo-halftone image through the conceivable error-distributing binary conversion process, however, the pseudo-halftone image will suffer from undesirable textures occurring in a leading edge S of the region R
3
. Pixels in leading edge S will be turned On at a regular interval. The thus regularly arranged turned-On pixels create a certain texture to be perceived by human eyes.
A similar phenomenon occurs also when the continuous tone image D is designed so that all the pixels in region R
2
have the maximum density (255) in the multilevel density range and so that all the pixels in region R
3
have the uniform density of a value lower than 255, for example, 254. When this continuous tone image D is converted into a pseudo-halftone image through the conceivable error-distributing image conversion process, pixels in leading edge S will be turned OFF also at a regular interval, thereby creating undesirable textures.
Accordingly, the pseudo-halftone image obtained based on the continuous tone image D has low quality.
In view of the above-described drawbacks, an object of the present invention is to provide an improved method for converting a continuous tone image into a pseudo-halftone image of high quality while preventing occurrence of undesirable textures in a transition region between a uniform density region having the minimum or maximum density and another uniform density region having a density (middle density) different from the corresponding minimum or maximum density.
In order to attain the above and other objects, the present invention provides a method for converting a continuous tone image into a pseudo-halftone image, the method comprising the steps of: receiving data of a continuous tone image desired to be converted into a pseudo-halftone image, the continuous tone image having a plurality of pixels, each pixel having one density value defined in a range between a predetermined minimum density and a predetermined maximum density; and subjecting all the pixels of the continuous tone image to an error-distributing binary conversion process on a pixel by pixel basis, to thereby convert the density values of the pixels into binary values while calculating conversion errors and distributing the calculated conversion errors to neighboring pixels, while performing at least one
Brinich Stephen
Brother Kogyo Kabushiki Kaisha
Lee Thomas D.
Oliff & Berridg,e PLC
LandOfFree
Error-distributing image conversion method does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Error-distributing image conversion method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Error-distributing image conversion method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2613243