JP3025714B2 - Pseudo gradation processor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a pseudo gradation processing apparatus for converting multi-gradation digital image data into binary image data having gradation, and is applied to a facsimile apparatus, an electronic filing apparatus and the like. It is suitable.

[0002]

2. Description of the Related Art As a pseudo-gradation display method for converting a multi-tone image into a binary image and perceiving density in a simulated manner by utilizing the integration of eyes, a density pattern in which one pixel is represented by a plurality of dots is used. A method and a dither method of expressing one pixel by one dot are known. FIG. 6 is a diagram showing the operation principle of the error diffusion method, which is a kind of the dither method. The error diffusion method performs binarization so as to minimize the density error between the original image and the result. The error data ε _ij stored in the error buffer memory 10 is weighted by the weighting coefficient generator 11. Coefficient α _ij
Are normalized and added by the adder 12 to the input multi-valued data x _ij . FIG. 7 shows an example of the weight coefficient α _ij .

The added data x _ij ′ output from the adder 12 is compared with a threshold value th by a binarizing circuit 13 to output binarized data y _ij . In the arithmetic unit 14, the addition data x
The difference between _ij ′ and the binarized data y _ij is output as error data ε _ij , and stored at the corresponding pixel position in the error buffer memory 10. By repeating this process, binarization of input multi-valued image data is executed.

[0004] The error diffusion method follows a change in the density of an original image well and has good compatibility between gradation and resolution, so that image quality evaluation is generally good.

[0005]

According to the above-described error diffusion method, a unique repetitive pattern called a texture is generated for a flat input with a small density change, and a linear pattern such as a character is generated. There is an inconvenience in that the edge portion has unevenness called a notch having a size of about one pixel.

SUMMARY OF THE INVENTION It is an object of the present invention to achieve both gradation and resolution of an output binary image, and to control fine structures such as textures, thereby reducing periodic textures and notches.

[0007]

A pseudo gradation processing apparatus according to the present invention converts input multi-gradation image data into binary image data.
A pseudo halftone processing unit for converting and outputting, is converted
An output buffer for storing the binary image data, attention
Binary pattern of pixel multi-tone values and peripheral output pixel positions
In advance, the black or white occurrence rate of the pixel of interest is
The recorded probability table and the input image data are
Multi-gradation of the input image data corresponding to the target pixel
The value and the image of interest stored in the output buffer.
Read the binary image data at the output pixel position around the element and
Refer to the probability table to determine the probability of black or white occurrence
Means for determining the probability of occurrence and the probability determined by the means for determining probability
Output pixel value that generates black or white binary image data at
Determining means.

[0008]

According to the present invention, when the multi-tone image is binarized and converted into a pseudo-tone image, the probability determining means determines whether the multi-tone image has already been binarized with the multi-tone value of the pixel of interest on the input multi-tone image. It has been backed
Of the output pixel position around the pixel of interest recorded in the
Probability of generating black or white for binary image data
Read from probability table, output pixel value determination means read
Then, black or white binary image data is generated with the given probability.

[0009]

[0010]

FIG. 1 is a block diagram showing an embodiment of a pseudo gradation processing apparatus according to the present invention, which comprises a probability determining unit 1, an output pixel value determining unit 2, and an output buffer 3. The probability determining unit 1 uses the pixel density at the target position on the input multi-tone image and the black-and-white arrangement pattern of binarized pixels near the pixel at the target position on the output binary image. The probability p that the pixel at the position of interest on the image is a black pixel is determined.

For example, 64 gradations (6 bits) as input
2 in the vicinity of the pixel of interest on the output binary image
Referring to the eight pixels marked with a cross as shown in FIG. 7, there are 256 (= 2 ⁸ ) black-and-white arrangement patterns, so that the pattern of the input density and the reference pixels is 16384 (= 64 × 2) in total.
56), and the probability p is determined by putting the probability of setting the target pixel in each case to black in the probability table.

The output pixel value determining unit 2 is a means for determining the value of the pixel of interest according to the probability p obtained by the probability determining unit 1. The probability is, for example, 10 pixels with a probability of 0.3
If there are, it means that three of them become black and seven become white. As described above, in order to determine the value of each pixel from the designated probability, the output pixel value determination unit 2 has the following configuration.

That is, from the random number generator 4, “0” to
A uniform random number in the range of “1” is generated, and the probability
Is compared with the output of the random number generation unit 4. If the random number is larger, the output pixel value is set to white, and the other values are set to black. Thus, a black pixel is generated based on the probability p. The output of the output pixel value determining unit 2 is stored in the output buffer 3 and fed back to the probability determining unit 1.

The nature of the binarization method by the pseudo gradation processing device largely depends on the contents of the probability table of the probability determining unit 1. In order to preserve the halftone, the probability p of the probability determination unit 1 may be set as a large tendency to be proportional to the input density. This is a binarization processing method called a random dither method. Furthermore, in this pseudo gradation processing device, since the binarized pixel arrangement pattern is used as information for determining the probability p, the texture of the output binary image can be controlled. That is, binarization processing of various properties can be realized by setting this table.

Here, it is considered that a sample image serving as a sample is set and the same properties are realized. For that purpose, the probability that an event corresponding to each entry of the table actually occurs on the sample image is measured, and this table is set using the value. First, a halftone image having a uniform density d and an image obtained by binarizing the halftone image by an error diffusion method are used as samples. This is called a first sample. Then, in the binary image of the first sample, in the case where the same eight pixels as in FIG. 2 are in the arrangement pattern X, the probability that the pixel at the target position is black is examined. Thus, the contents of the entry of the input density d and the output pattern X in the table are obtained. This is repeated for all densities and patterns to determine the contents of the table.

In a sample image, it never appears,
Alternatively, there is an entry whose appearance frequency is too low and the reliability of the obtained probability is low, but this can be arbitrarily determined as an exception process. For example, the input density d is directly set as the probability p regardless of the output pattern. In this way, the table can be set. When the binarization processing according to the present invention is performed using this table, a binary image having the same texture as the sample error diffusion image can be obtained for a uniform density or a gradual density change input.

Even if a table in which the first sample is set by measurement is used, if the sample binary image has a periodic texture, the periodic texture is likely to be generated. Therefore, as the second sample, the same uniform density image as the first sample is taken as the multi-tone image.
A small random noise is added to the uniform-density image as a binary image, and then a binary image without periodic texture is obtained by binarization by an error diffusion method. When the probability table set by the second sample is used, the periodicity is further reduced.

Another sample can be taken.
As shown in FIG. 4A, an image having both a photographic image portion having a gradual change in density and a character image portion having a sharp change is taken as a multi-tone image. On the other hand, as shown in FIG. 4B, a photographic image portion is binarized by an error diffusion method as a sample binary image, and a character image portion is binarized by fixed threshold processing after MTF correction. take. This is called a third sample.

The probability table obtained at this time is different from that when the second sample is used. In particular, a difference appears in the portion as shown in FIG. In the second sample, such a pattern hardly appears. Therefore, the input density itself is used as a probability by the above-described exception processing. However, in the third sample, such patterns frequently appear at the edge portion of the character, and the pixel of interest has a high probability of being black. That is, in the present method using such a probability table, a notch is small in an edge portion, and a binary image similar to that in the error diffusion process can be obtained in other portions than the edge portion.

[0020]

Effect of the Invention] The present invention, to determine the probability of the output pixel value from the input pixel density and the output pixel pattern, is determined
Output black or white binary image data
To reduce textures and notches.
it can.

[0021]

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention.

FIG. 2 is a diagram illustrating reference pixels of an output binary image.

FIG. 3 is a block diagram of an output pixel value determining unit.

FIG. 4 is a diagram showing a third sample image.

FIG. 5 is a diagram showing an edge image.

FIG. 6 is an operation principle diagram of the error diffusion method.

FIG. 7 is a diagram illustrating an example of a weight coefficient.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Probability determination part 2 Output pixel value determination part 3 Output buffer 4 Random number generation part 5 Comparison part

Claims (2)

(57) [Claims] An input multi-tone image data is converted to a binary image data.
A pseudo halftone processing unit for converting and outputting the output buffer for storing the converted the binary image data
And the binary value of the multi-tone value of the pixel of interest and the peripheral output pixel position.
Black or white occurrence of the target pixel in advance corresponding to the turn
A probability table in which the rate is recorded and the input image data are made to correspond to the pixel of interest, and
Multi-tone value of input image data and output buffer
Of the output pixel position around the pixel of interest stored in
Reading the binary image data and referring to the probability table,
A probability determining means for determining the probability of occurrence of black or white; and a binary image of black or white determined by the probability determined by the probability determining means.
A pseudo-grayscale processing device comprising: output pixel value determining means for generating image data . 2. The method according to claim 1, wherein said output pixel value determining means is in a range of 0 to 1.
A uniform random number is generated and determined by the probability determining means.
Black or white binary image data is generated
The pseudo gradation processing device according to claim 1 , wherein the pseudo gradation processing device is configured to perform the processing.