Image analysis – Image compression or coding – Gray level to binary coding
Reexamination Certificate
1998-11-10
2002-05-14
Lee, Thomas D. (Department: 2624)
Image analysis
Image compression or coding
Gray level to binary coding
C382S270000
Reexamination Certificate
active
06389172
ABSTRACT:
BACKGROUND OF THE INVENTION
1. [Field of the Invention]
The present invention relates to an image processing apparatus for converting input image data into black/white binary image data and, a storage medium that stores a program associated with the apparatus.
2. [Description of the Related Art]
Binarization for converting a halftone image into a black/white binary image is used not only in processing of image data (e.g., texture source images in game machines or the like) but also in every field of image signal processing such as facsimile, OCR (optical character reader), medical image processing, visual robot, and the like. As a conventionally popular binarization method, a pseudo halftone display method based on so-called “ordered” dithering, which superposes random numbers or pseudo random numbers (periodic random numbers) on a halftone image, and then A/D (analog/digital)-converts the halftone image, is known. As a method of giving random numbers or pseudo random numbers used, a Bayer pattern is preferably used.
FIG. 4A
shows the principle of binarization based on “ordered” dithering.
FIG. 4B
shows an example of a halftone image to be input to the binarization shown in
FIG. 4A. A
halftone image
420
shown in
FIG. 4B
consists of a set of 16(=4×4) pixels (dots)
421
,
422
, . . . The numeral in the box of each pixel represents the gradation value of that pixel. For example, “8” is described in the box of the pixel
421
, and “7” is described in the box of the pixel
422
. These values respectively represent the gradation values of the corresponding pixels. In this case, the gradation level of each pixel is expressed by a numerical value ranging from 0 to 15 (i.e., 16 gradation levels).
In the binarization shown in
FIG. 4A
, when the halftone image
420
is input, a pseudo random number generator generates a pseudo random number (Bayer) pattern having 16(=4×4) elements, and a subtracter
402
subtracts the corresponding element values of a pseudo random number pattern
401
from the gradation values of the individual pixels of the halftone image
420
. For example, a random number “0” of an element
411
in the pseudo random number pattern
401
is subtracted from the gradation value “8” of the pixel
421
in the halftone image
420
, and subtractions between the corresponding pixels and elements are similarly done. A discrimination circuit
403
outputs a dither image as a result of binarizing an element equal to or larger than 0 (or an element that exceeds 0) to 1 (means white) and an element that does not exceed 0 (or an element equal to or smaller than 0) to 0 (means black) in the subtraction results from the subtracter
402
. In this case, 1 indicates white, and 0 black. But 1 may indicate black, and 0 white. The same applies to the embodiment of the present invention to be described later.
In the description of
FIG. 4A
, binarization is done by the subtracter
402
and discrimination circuit
403
. To summarize, this processing compares the gradation values of the individual pixels in the halftone image
420
with the element values (to be referred to as threshold values hereinafter) in the pseudo random number pattern
401
, and sets 1 (white) for a pixel at which the gradation value is equal to or larger than the threshold value) or 0 (black) for a pixel at which the gradation value does not exceed (or is equal to or smaller than) the threshold value. Such binarization may be realized by hardware or implemented by software.
When an input halftone image is large, it is broken up into ranges of 16 pixels(=4×4), as shown in
FIG. 4C
, and individual ranges
431
,
432
, . . . are binarized by the aforementioned method to obtain a final dither image. In this case, two-dimensional blocks each consisting of 16(=4×4) pixels (i.e., the order of pattern=4) are used as units.
Alternatively, two-dimensional blocks having other sizes, e.g., 64(=8×8) pixels (the order of pattern=8), 256(=16×16) pixels (the order of pattern=16), and the like may be used as processing units.
FIG. 5A
shows recurrence formulas used for generating the pseudo random number pattern used in the aforementioned binarization, i.e., a Bayer pattern. For example, if n=2 in
FIG. 5A
, a pattern of “D
4
” (the order of pattern=4) (i.e., the pseudo random number pattern
401
shown in
FIG. 4A
) is obtained. In normal use, since the number of gradation levels of an input halftone image to be binarized is 256 (a gradation value ranges from 0 to 255) expressed by 8 bits, a pattern
501
(
FIG. 5B
) obtained by multiplying each threshold value in the pattern
401
by 17(=255/15) is popularly used.
When already binarized data (e.g., a binary image, each pixel of which does not assume a value other than 0 indicating black or
255
indicating white) in place of a halftone image, periodic noise components appear in the output image. This will be explained below.
Assume that binarization is done using a pseudo random number pattern
501
shown in FIG.
5
B. If the gradation values of the respective pixels of the input image are compared with the individual threshold values in the pattern
501
using:
pixel that satisfies gradation value≧threshold value→white
pixel that satisfies gradation value<threshold value→black (i)
since gradation value≧threshold value always holds for elements with a threshold value=0 in the pattern
501
, the corresponding pixel of the dither image always becomes white. As a consequence, if the input image is already binarized, white isolated pixels periodically appear, as indicated by
601
in FIG.
6
. Note that
601
in
FIG. 6
indicates dither image blocks output when an image, all the pixels of which are black (gradation value=0) is input using the pseudo random number pattern
501
. Since the upper left threshold value in the pattern
501
is 0, a pixel at that position in each output dither image block
601
is white.
On the other hand, if the gradation values of the respective pixels of the input image are compared with the individual threshold values in the pattern
501
using:
pixel that satisfies gradation value>threshold value→white
pixel that satisfies gradation value≦threshold value→black (ii)
since gradation value≦threshold value always holds for elements with a threshold value=255 in the pattern
501
, the corresponding pixel of the dither image always becomes black. As a consequence, if the input image is already binarized, black isolated pixels periodically appear, as indicated by
602
in FIG.
6
. Note that
602
in
FIG. 6
indicates dither image blocks output when an image, all the pixels of which are white (gradation value=255) is input using the pseudo random number pattern
501
. Since the lower left threshold value in the pattern
501
is 255, a pixel at that position in each output dither image block
602
is black.
If a method of comparing an input image with a fixed threshold value (e.g., 128) is used, such problems are not posed. However, a binary image obtained from an input halftone image by such method cannot be used.
SUMMARY OF THE INVENTION
The present invention has been made in consideration of the conventional problems, and has as its object to provide an image processing apparatus, which can prevent noise, i.e., periodic appearance of white or black isolated pixels even when an already binarized image is input in binarization that normally receives a halftone image, and can obtain a dither image having quality as high as that obtained by the conventional method even when a halftone image is input.
In order to achieve the above object, an image processing apparatus of claim
1
comprises input means for inputting digital image data, pseudo random number pattern generation means for generating a pseudo random number pattern which consists of, as elements, threshold values set at substantially equal intervals without contain
Brinich Stephen
Lee Thomas D.
Morrison & Foerster / LLP
Namco Limited
LandOfFree
Image processing apparatus and storage medium that stores... 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 storage medium that stores..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing apparatus and storage medium that stores... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2878755