JP3480924B2 - Image forming method, image forming apparatus, and recording medium - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image forming method applied to an image forming apparatus such as a laser printer, a digital copying machine, a color laser printer and a digital color copying machine, and a recording recording an image forming processing program. Regarding the medium.

[0002]

2. Description of the Related Art An error diffusion method is known as a pseudo halftone process for expressing a halftone with binary dots. This method is a method of diffusing an error that occurs when binarizing each pixel to peripheral pixels. For example, by outputting a different binarization threshold level depending on the peripheral dot arrangement after binarization,
A method of forming an image with good gradation and sharpness by ensuring the continuity of figures (see Japanese Patent Publication No. 8-24338), and when dots are present in peripheral pixels after binarization at low density. There is a method (see Japanese Patent No. 2662402) for preventing the generation of texture by connecting dots in a low density portion by turning off output dots.

Further, there is a multi-valued error diffusion method as an extension of this error diffusion method to three or more values. For example, the quantization number of error diffusion is switched according to the image characteristics near the pixel of interest (that is, the average value of the quantized peripheral data and the edge amount of the image). In the tonal image section, a method of making the sharpness of characters and the graininess of a halftone image compatible by increasing the number of multi-valued levels (see Japanese Patent No. 2851662), a threshold value and a quantization level depending on the input data level. A variation range is set, and a threshold value and a quantization level are determined within the variation range. A method of varying the threshold value and the quantization level for the purpose of preventing false contours (see Japanese Patent Laid-Open No. 11-252364) is a method. is there.

In the case of binary error diffusion, dot ON / OFF
Since the gradation is expressed by F, the stability of the image is excellent, but there is a problem that the isolated dots are conspicuous in the low density portion and the granularity is poor. On the other hand, in the multi-valued error diffusion, small dots are output in the low density portion, so isolated dots are less noticeable and have good graininess.

[0005]

However, there is a problem that when the density level becomes high, the number of dots at the small level increases and the image becomes unstable. In addition, due to the increase in the number of dots of a small level, the saturation of density is quick and the number of gradations is reduced. Such a problem is particularly remarkable in image formation by an electrophotographic printer.

An object of the present invention is to eliminate the instability in the middle and high density areas while maintaining the graininess in the low density areas, which is an advantage of multi-valued error diffusion, and to provide an image forming method and image with good gradation. To provide a forming apparatus and a recording medium.

[0007]

According to the present invention, the threshold value is changed according to the input data level, small level dots are output in the low density area of the image, and small level dots are generated in the middle and high density areas of the image. By reducing the probability, both the graininess of the low density portion of the image and the stability of the middle and high density portions of the image are achieved.

According to the present invention, by changing the error diffusion threshold according to the multi-tone image data level, the probability of small dots occurring in the middle and high density areas of the image is lowered, and the image stability is improved. To do.

According to the present invention, the threshold value is controlled in consideration of the peripheral data by changing the threshold value according to the value obtained from the data level of the pixel of interest and its periphery.

According to the present invention, as the data level increases, the interval between the two threshold values that determine the output of small dots is narrowed to reduce the probability of small dots occurring in the middle and high density areas of the image.

According to the present invention, the two thresholds are matched at a predetermined data level or higher so that small dots are not generated in the middle and high density portions of the image (that is, the number of quantization levels is reduced).

According to the present invention, gradation jump is prevented by changing the threshold value stepwise.

According to the present invention, by selecting a fixed threshold value based on the image feature and a threshold value that changes according to the data level (hereinafter referred to as “variable threshold value”), it is possible to obtain an optimum small dot generation ratio for the image feature. Control to be.

According to the present invention, an image with good sharpness is formed by selecting a fixed threshold value in the image edge portion, and a variable threshold value is selected in the non-edge portion of the image, resulting in good graininess in the low density portion.・ Forms an image with good stability in high-density areas.

In the present invention, a fixed threshold value is selected in the character portion of the image to form an image with good sharpness, and a variable threshold value is selected in the image portion of the image to obtain good graininess in the low density portion,・ Forms an image with good stability in high-density areas.

According to the present invention, the degree of change of the variation threshold is changed stepwise according to the characteristics of the image, so that an image having no discomfort at the change point of the characteristics of the image is formed.

According to the present invention, the edge degree of the image is increased. Therefore, by reducing the degree of change of the variation threshold, an image with good sharpness is formed at the edge portion of the image, and switching to the non-edge portion is performed. To form a comfortable image.

Further, in the present invention, the degree of change of the variation threshold is switched according to the output mode, so that the generation ratio of small dots is optimized for each output mode.

[0019]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be specifically described below with reference to the drawings. FIG. 19 shows an example of an image forming apparatus including a digital copying machine to which the present invention is applied.
This digital copying machine includes a scanner 400 as an image reading device, an image recording device 411 including a laser printer as an image forming unit, and a circuit described later.
The scanner 400 illuminates an original such as a bookbinding original placed on a flat original table 403 with an illumination lamp 502,
The reflected light image is reflected by the mirror groups 503 to 505 and the lens 5.
An image is formed on the reading sensor 507 via 06, and the document is scanned by the movement of the illumination lamp 502 and the mirror groups 503 to 505 to read the image information of the document and convert it into an electrical image signal. The image signal obtained by the reading sensor 507 is transferred to the printer 41 via a circuit described later.
Sent to 1.

In the printer 411, a writing device 508 including a writing optical unit as an exposing unit.
Forms an electrostatic latent image by converting the image signal into an optical signal and exposing the image bearing member made of a photoconductor, for example, the photoconductor drum 509, to perform optical writing corresponding to the original image. The writing optical unit 508 drives a semiconductor laser with the image signal by the light emission drive control unit to emit laser light whose intensity is modulated by the image signal, and deflects and scans this laser light with the rotary polygon mirror 510 to generate f / θ. The photosensitive drum 509 is irradiated with light through the lens and the reflection mirror 511.

The photosensitive drum 509 is rotationally driven by a driving unit to rotate clockwise as indicated by an arrow, is uniformly charged by a charger 512 as a charging means, and is then electrostatically exposed by the writing optical unit 508. A latent image is formed. The electrostatic latent image on the photoconductor drum 509 is developed by the developing device 513 into a toner image, and the transfer material made of transfer paper is one of a plurality of paper feed units 514 to 518 and a manual paper feed unit 519. Is fed to the registration roller 520.

The registration roller 520 is the photosensitive drum 50.
The transfer sheet is sent in time with the toner image on the transfer sheet 9, and a transfer bias is applied to the transfer belt 521 from the transfer power source to convey the transfer sheet, and at the same time, the photosensitive drum 50
The toner image on 9 is transferred to a transfer paper. The transfer sheet is conveyed by the conveyance belt 521, the toner image is fixed by the fixing unit 522, and the transfer sheet is ejected to the ejection tray 523 as a copy. In addition, the photosensitive drum 509 is cleaned by the cleaning device 524 after the toner image is transferred, and is gradually discharged by the discharging device 525 to prepare for the next image forming operation.

FIG. 1 shows the configuration of the image processing section of the digital copying machine. The image processing unit includes a scanner system processing unit that corrects read data, a digital image processing unit that processes and corrects a digital image, and a writing system processing unit that modulates a writing LD. The analog data of 600 dpi read by the CCD1 is adjusted in data level by the AGC2. Next, the AD conversion 3 converts the analog data of each pixel into a digital value of 8 bits per pixel, and the shading correction 4 corrects the variation of the pixels of the read CCD and the illuminance.

Subsequently, filter processing 5 is performed. Specifically, MT for correcting the amplitude of an image generated by reading
F correction and smoothing processing for smoothly expressing a halftone image are performed. Then, a scaling process 6 in the main scanning direction is performed according to the copy magnification, and a γ correction 7 for converting to a writing density is performed. Finally, a halftone process 8 is performed, and the data is converted into data of several bits per dot and transmitted. In addition, background removal processing, flare removal processing, scanner γ processing, image editing, and other processing (not shown) are performed.

Next, the image separation processing section will be described.
Generally, the input image is a mixture of characters, halftone dots, and photographic images. In the character image, black and white pixels are continuous at the edge portion of the image, and peaks periodically exist in the halftone image (pattern). The feature determining unit 10 detects each of these image features to convert each pixel of the input image into a character 11 and a photograph 1.
2 or halftone dot 13 can be separated.

This separation information is sent to the halftone processing section 8 and is used for switching the halftone processing and the processing parameters.

The error diffusion processing for quantizing the image data according to the present invention is provided in the halftone processing section 8.

FIG. 2 shows the configuration of a conventional error diffusion process. The error diffusion process compares the multi-tone image data 20 with a threshold value 22 to determine an output value 23. And the output value 2
The difference 24 between 3 and the image data 20 is stored as the error 25 of the pixel. At the next pixel, the product of the peripheral pixel and the error matrix coefficient 26 corresponding to the pixel is added to the image data of the pixel of interest 21, and the output value 2 of the pixel is compared with the threshold 22.
Determine 3. By repeating the above for each pixel,
An error diffusion process is performed in which the image density is saved.

FIG. 3 shows the ratio of output dots in the ternary error diffusion process as an example of the multi-valued error diffusion process. Output value for each of the quantization levels 0 to 2 is 0 (dot OF
F), 128 (for example, small size dot), and 255 (for example, large size dot), the ratio of the quantization level 1 increases until the input data level is 128 (the number of small size dots increases as the density increases). Increase), the ratio of the quantization level 1 at the input data level 128 becomes 100%.
When the input data level is 128 to 255, the ratio of the quantization level 1 decreases and the ratio of the quantization level 2 increases (the number of small size dots decreases and the number of large size dots increases as the density increases). , Input data level 255
Then, the ratio of the quantization level 2 becomes 100%. As described above, in multi-valued error diffusion, there is a region in which the proportion of small level dots is high, and there is a problem that the image becomes unstable in that region.

In the present invention, this problem is solved and a high quality image is formed by controlling the generation rate of small dots.

(Embodiment 1) FIG. 4 shows the configuration of Embodiment 1 of the present invention. A threshold value generation unit 27 is further added to the configuration of the error diffusion processing of FIG.

The first embodiment will be described below for the case of quantizing into three values. The threshold value generating unit 27 of FIG. 4 uses the threshold value table that changes according to the input data as shown in FIG. 5, and uses the two threshold values Thr1 and Thr corresponding to the input data.
Determine 2. Quantization is performed using the thresholds Thr1 and Thr2.

The total value of the input data and the marginal error (hereinafter, correction input value) is compared with the threshold value, and when the correction input value <Thr1, the output = 0 (quantization level 0) Thr1 ≦ the correction input value <Thr2, the output = 128 (quantization level 1) Thr2 ≦ correction input value Output = 255 (quantization level 2)

When the input data is less than A, Thr1 and Th
Since the value of r2 is different, the correction input value is quantized into three values, and the dot of the quantization level 1 appears, but if the input data is A or more, the values of Thr1 and Thr2 are the same, so the correction input value is It is quantized into a binary value, and a dot of quantization level 1 does not appear. When the input data is less than A, Thr1 and Thr become larger as the input data becomes larger.
Since the difference between 2 becomes smaller, the output of the quantization level 1 dot is suppressed as the input data becomes closer to A, and when the input data becomes A, the quantization level 1 dot is not output. Therefore, since the generation rate of each quantization level is smoothly switched with respect to the gradation change of the input image, gradation skipping or pseudo contour does not occur.

FIG. 8 shows the dot generation rate in this embodiment. With input data of a certain level (B) or higher, the ratio of the dots of the quantization level 1 decreases, and the dots of the quantization level 2 are output. The input data A and above are quantized into binary values of quantization level 0 and quantization level 2.

Therefore, as shown in FIG. 3, since the ratio of the dots of the quantization level 1 does not become 100%, it is possible to prevent the image from becoming unstable in the middle and high density portions, and the dots can be prevented. Gradation saturation due to gain can also be suppressed.

By setting the setting level of A to, for example, 60, the maximum generation rate of the dots of the quantization level 1 can be suppressed to a relatively small value. In order to prevent the generation rate of the quantization level 1 from becoming 100%, it is necessary that the output value of the quantization level 1 (128 in this embodiment) is less than that.

Similar effects can be obtained even if the threshold value tables used in the threshold value generating section 27 are those shown in FIGS. In FIG. 6, the value of Thr1 is not changed and Th
The value of r2 is changed so as to become the same value as Thr1 when A or more. By not changing the value of Thr1 in this way, it is possible to prevent a dot generation delay in the low density portion of the image.

Further, as shown in FIG. 7, when the input data is low, the values of Thr1 and Thr2 do not change, and when A or more, both Thr1 and Thr2 are set to 128, and even if a threshold table is used, dots in the low density portion of the image are It is possible to prevent the occurrence delay of.

(Embodiment 2) FIG. 9 shows the configuration of Embodiment 2 of the present invention. The difference from the first embodiment is that an averaging unit 28 is provided, and instead of determining the threshold value according to the value of the input data of the target pixel, in the present embodiment, the average value of the input data of the target pixel and its peripheral pixels is used. To determine the threshold.

The peripheral pixels are, for example, 3 × 3 pixels centered on the pixel of interest (*) as shown in FIG. 10, and are quantized by a threshold table that changes according to the average value of the multi-tone input data of the 9 pixels. Determine the activation threshold. As the threshold table, the tables of FIGS. 5, 6 and 7 are used as in the first embodiment. However, in this case, the horizontal axis of FIGS. 5, 6 and 7 is the average value instead of the input data.

The ratio and effect of dot generation in this embodiment are similar to those in the first embodiment, but by taking the average value with the surrounding pixels, the influence of noise in the input image can be reduced and appropriate threshold control can be performed. Is.

Further, instead of calculating the simple average of the pixel of interest and peripheral pixel data as described above, it is also possible to calculate a weighted average or a moving average, or use a value calculated by filtering (for example, smoothing). Is.

(Third Embodiment) FIG. 11 shows a third embodiment of the present invention.
Shows the configuration of. In the present embodiment, a feature amount detection unit 33 and a selection unit 34 are added to the configuration of the first embodiment. A plurality of threshold value generating units 29 to 32 configured by a threshold value table are provided, and the selecting unit 34 switches the threshold value table (threshold value generating unit) according to the feature amount of the image and uses the threshold value table to correspond to the input data. Two thresholds Thr1 and Thr2
To decide. The present embodiment will be described below in the case where the edge amount of the image is used as the feature amount of the image.

The calculation of the edge amount of the image is performed from FIG.
The edge amount in each direction is calculated using the differential filter of 4 × 5 × 5 pixels shown in (d), and the maximum value among the edge amounts in the four directions is set as the edge amount of that pixel. Four levels of edge levels are determined based on this edge amount. At the edge level, the edge level 0 is assigned so as to become a flat portion of an image having a small edge degree, and the edge level 3 is assigned as an image edge portion having a large edge degree.

According to the edge level thus determined, the selecting unit 34 selects the threshold value table (threshold value generating units 29 to 32) for determining the quantization threshold value. Four types of threshold tables shown in FIGS. 13 to 16 are prepared. FIG. 13 is used for edge level 0 (flat image portion) and FIG. 1 is used for edge level 1.
4 is selected for edge level 2, and FIG. 16 is selected for edge level 3 (image edge portion). That is, the larger the edge degree, the smaller the fluctuation of the threshold value with respect to the input data, and at the maximum level of the edge degree, the fixed threshold value does not change with respect to the input data. Two thresholds Thr corresponding to the input data are selected using the selected threshold table.
1, Thr2 are determined, and the same quantization as in the first embodiment is performed.

By switching the threshold value table according to the edge degree of the image as in the present embodiment, since the error diffusion of the fixed threshold value is performed at the edge portion of the image, an image with good sharpness is formed. In the flat part of the image,
Since the generation rate of small dots is reduced in the high density portion, it is possible to prevent the image from becoming unstable, and it is also possible to suppress gradation saturation due to the dot gain. Furthermore, since the edge level is set to four levels, the generation ratio of small dots is switched stepwise between the edge part and the flat part of the image, and an image with no discomfort at the boundary between the edge part and the flat part can be formed.

In the present embodiment, in addition to the above edge amount as the image feature amount, it is also possible to use image information such as photographs, halftone dots, and characters that have undergone image separation processing in FIG.
In this case, for example, by selecting the threshold table of FIG. 13 for the photograph area, FIG. 15 for the halftone area, and FIG. 16 for the character area, the stability of the middle and high density areas and the sharpness of the character area in the photograph area are selected. It is possible to achieve both sex.

Further, in this embodiment, as in the second embodiment, it is possible to use the average value of the input data of the target pixel and its peripheral pixels instead of the value of the input data of the target pixel.

(Fourth Embodiment) FIG. 17 shows a fourth embodiment of the present invention.
Shows the configuration of. The difference from the third embodiment is that an output mode designating unit 35 is provided instead of the feature amount detecting unit, and the selecting unit 34 switches the threshold value table according to the designated output mode. The threshold table (threshold generation units 29 to 32) is switched depending on the output mode, and the two thresholds Thr1 corresponding to the input data are used by using the threshold table.
Determine Thr2.

The output mode designating section 35 is instructed by a user from an operation panel or the like.
Photo mode suitable for smooth gradation reproduction of photos etc., character mode suitable for reproducing density change points of characters etc. at high resolution, character / photo mode optimal for originals with mixed photographs and characters, printed matter etc. There is a print photo mode most suitable for halftone originals.

According to this output mode, the selecting section 34 selects the threshold table (threshold generating sections 29 to 32). For example, FIG. 13 in the photo mode, FIG. 16 in the character mode, FIG. 14 in the character / photo mode, and FIG. 15 in the print photo mode.
By selecting the threshold table of, it is possible to suppress the generation of small dots in the middle and high density areas in the photo mode to form a stable image, and to form a highly sharp image by the error diffusion of the fixed threshold in the character mode. You can

Also in this embodiment, as in the second embodiment, it is possible to use the average value of the input data of the target pixel and its peripheral pixels instead of the value of the input data of the target pixel.

(Fifth Embodiment) The first to fourth embodiments described above.
In the above, the case of quantizing into three values has been described, but it is also possible to quantize into four values or more by the same method. FIG. 18 shows an example of a threshold value table when quantizing into four values. By using such a threshold value table, the input data of 50 or less is quantized into four values, the input data of 50 or more and less than 150 is quantized, and the input data of 150 or more is quantized into binary.

Therefore, the quantization level 1 in the middle density part
Dots and dots of the quantization level 2 in the high density portion are suppressed, and an image with good stability can be formed.
In addition, when the quantization level 2 dots are not suppressed,
Thr3 and Thr2 even when the input data is over 150
A threshold table that does not match (= Thr1) may be used.

The output values of the parameters, threshold values, and quantization levels described in each of the above embodiments have different optimum values depending on the output device, so the present invention is limited to the set values described in each embodiment. is not.

The present invention can also be realized by software. A processing procedure for performing image formation or a program for executing a processing function of the present invention is recorded in a computer-readable recording medium or the like. When executing the image forming processing of the present invention in a digital copying machine or the like, the program is installed from the above medium into the system of the digital copying machine, and the processing is started.

[0058]

As described above, according to the present invention, the following effects can be obtained. (1) By changing the error diffusion threshold in accordance with the multi-tone image data level, the probability of small dots occurring in the middle and high density areas of the image can be lowered, and an image with good stability can be formed. it can. (2) As the data level increases, the interval between the two thresholds that determine the output of small dots is narrowed, so that the probability of small dots occurring in the middle and high density portions of the image can be reduced. (3) At the predetermined data level or higher, the two thresholds are matched so that small dots are not generated in the middle and high density portions of the image. (4) Gradation jump can be prevented by changing the threshold value stepwise. (5) The influence of image noise can be reduced by changing the threshold value according to the average value of the data level of the pixel of interest and its surroundings. (6) It is possible to reproduce an image with an optimal small dot generation rate for the characteristics of the image. (7) By selecting a fixed threshold value in the image edge portion, an image with good sharpness is formed, and by selecting a variable threshold value in the non-edge portion of the image, the low density portion has good graininess, and the medium / high density portion Thus, an image with good stability can be formed. (8) By selecting a fixed threshold value in the character portion of the image, an image with good sharpness is formed, and in the design portion of the image, a variable threshold value is selected so that the low density portion has good graininess and the medium / high density portion Thus, an image with good stability can be formed. (9) By gradually changing the degree of change in the variation threshold according to the characteristics of the image, it is possible to form an image that does not cause a sense of discomfort at the change point of the characteristics of the image. (10) As the degree of edge of the image increases, the degree of change of the variation threshold is reduced to form an image with good sharpness at the edge portion of the image, and it does not cause discomfort due to switching to the non-edge portion. An image can be formed. (11) By changing the degree of change in the variation threshold according to the output mode, it is possible to reproduce an image with an optimum small dot generation rate for each output mode.

[Brief description of drawings]

FIG. 1 shows a configuration of an image processing unit of a digital copying machine.

FIG. 2 shows a configuration of a conventional error diffusion process.

FIG. 3 shows a ratio of output dots in a conventional ternary error diffusion process.

FIG. 4 shows a configuration of a first embodiment of the present invention.

FIG. 5 shows a first example of a threshold table.

FIG. 6 shows a second example of a threshold table.

FIG. 7 shows a third example of a threshold table.

FIG. 8 shows a ratio of output dots in the first embodiment.

FIG. 9 shows a configuration of a second embodiment of the present invention.

FIG. 10 shows an example of a peripheral pixel area.

FIG. 11 shows a configuration of Example 3 of the present invention.

12A to 12D show edge extraction filters.

FIG. 13 shows a first example of a threshold table selected according to the edge amount.

FIG. 14 shows a second example of a threshold table selected according to the edge amount.

FIG. 15 shows a third example of a threshold table selected according to the edge amount.

FIG. 16 shows a fourth example of a threshold table selected according to the edge amount.

FIG. 17 shows a configuration of a fourth embodiment of the present invention.

FIG. 18 shows an example of a four-valued threshold table.

FIG. 19 shows an example of an image forming apparatus including a digital copying machine to which the present invention is applied.

[Explanation of symbols]

1 CCD 2 AGC 3 AD converter 4 Shading correction section 5 Filter processing unit 6 Main scanning zoom unit 7 γ correction unit 8 Halftone processing section 9 LD 10 Feature determination unit 11 character part 12 Photo Department 13 halftone dot 20 input data 21 adder 22 Quantizer 23 Output data 24 Error calculator 25 error buffer 26 Error diffusion matrix 27 Threshold generator

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) H04N 1/40 B41J 2/52 G06T 5/00 200

Claims (22)

(57) [Claims] 1. An image forming method for quantizing input data of an m-valued multi-gradation image into an n value (3 ≦ n <m) by an error diffusion method, wherein the input data is equal to or higher than a predetermined level. An image forming method, characterized in that a quantization threshold is changed according to the level of the input data so that the probability of occurrence of small dots becomes low. 2. As the input data increases,
2. The image forming method according to claim 1, wherein the difference between two quantization threshold values that determine the output of small dots is reduced. 3. The image forming method according to claim 2, wherein the two quantization thresholds are matched when the input data is equal to or higher than a predetermined level. 4. The image forming method according to claim 2, wherein the difference between the quantization thresholds is reduced stepwise. 5. An image forming method for quantizing input data of an m-valued multi-gradation image into an n-valued (3 ≦ n <m) by an error diffusion method, which is a pixel of interest of the input data and its peripheral pixels. Multi-valued data is obtained from the data, and when the multi-valued data is above a predetermined level, the quantization threshold is changed according to the level of the multi-valued data so that the occurrence probability of small dots becomes low. Image forming method. 6. The multi-valued data is an average value of data of a target pixel and its peripheral pixels.
The image forming method described. 7. An image forming method for quantizing input data of an m-valued multi-gradation image into an n-value (3 ≦ n <m) by an error diffusion method, which is changed according to the level of the input data. An image forming method, characterized in that the degree of change in the threshold value is changed according to the output mode. 8. The output mode includes a photo mode, a character / photo mode in which photographs and characters are mixed, a print photo mode, and a character mode, and the photo mode, character / photo mode,
The image forming method according to claim 7 , wherein the degree of change in the threshold value is reduced in the order of the print photograph mode and the character mode. 9. When the output mode is the character mode, Motomeko 7 or 8 image forming method according to, characterized in that the threshold fixed. 10. An image forming apparatus comprising a quantizing means for quantizing input data of an m-valued multi-tone image into an n-valued (3 ≦ n <m) by an error diffusion method, wherein the input data A threshold value generating unit that generates a threshold value that changes according to the threshold value; and when the input data is equal to or higher than a predetermined level, a quantum value that is generated by the threshold value generating unit is quantized so that the probability of occurrence of small dots becomes low. And an image forming device. Wherein said threshold generating means, according to claim 1, wherein the input data in accordance with increases, to reduce the difference between two quantization thresholds that determines the output of the small dots
The image forming apparatus according to item 0 . 12. The image forming apparatus according to claim 10 , wherein the threshold value generation unit matches the two quantization threshold values when the input data is equal to or higher than a predetermined level. Wherein said threshold generating means, according to claim, characterized in that to reduce the difference in the quantization threshold stepwise 10
The image forming apparatus described. 14. An image forming apparatus comprising a quantizing means for quantizing input data of an m-valued multi-gradation image into an n-value (3 ≦ n <m) by an error diffusion method. Multi-valued data calculation means for calculating multi-valued data from the data of the target pixel and its peripheral pixels, threshold value generation means for generating a threshold value that changes according to the multi-valued data, and the multi-valued data having a predetermined level or more. At this time, the image forming apparatus is provided with a quantizing unit that quantizes by using the threshold value generated by the threshold value generating unit so that the generation probability of the small dots becomes low. 15. The image forming apparatus according to claim 14, wherein the multi-valued data calculation unit calculates an average value of data of the pixel of interest and its peripheral pixels. 16. An image forming apparatus comprising a quantizing means for quantizing input data of an m-valued multi-gradation image to an n-valued (3 ≦ n <m) by an error diffusion method, wherein the input data A feature amount detecting means for detecting a feature amount, a plurality of threshold value generating means for generating a threshold value that changes according to the input data, and a plurality of threshold value varying means having different degrees of change in the threshold value, and a plurality of the plurality of threshold value generating means based on the detected feature amount. An image forming apparatus comprising: a selecting means for selecting one threshold generating means from the threshold generating means, and a quantizing means for quantizing using the threshold generated by the selected threshold generating means. . 17. The image forming apparatus according to claim 16 , wherein when the feature quantity is an edge portion, the selecting means selects a threshold value generating means for generating a fixed threshold value. 18. The image forming apparatus according to claim 16 , wherein when the characteristic amount is a character portion, the selecting unit selects a threshold generating unit that generates a fixed threshold. 19. The method according to claim 16 , wherein when the feature amount is an edge amount, the selecting unit selects a threshold value generating unit that reduces the degree of change of the threshold value as the edge degree increases. Image forming device. 20. An image forming apparatus comprising a quantizing means for quantizing input data of an m-valued multi-gradation image into an n-valued (3 ≦ n <m) by an error diffusion method, wherein an output mode is designated. Output mode designating means for generating a threshold value that changes according to the input data, a plurality of threshold value generating means having different degrees of change in the threshold value, and a plurality of threshold value generating means based on the designated output mode. An image forming apparatus comprising: a selection unit that selects one threshold value generation unit from among them, and a quantization unit that quantizes using a threshold value generated by the selected threshold value generation unit. 21. A computer read recording program for causing a computer to realize an image forming processing function of quantizing input data of an m-valued multi-gradation image into an n-valued (3 ≦ n <m) by an error diffusion method. The recording medium is capable of generating a threshold value that changes according to the input data, and when the input data is at a predetermined level or higher, the threshold value that is generated is set so that the small dot occurrence probability becomes low. A computer-readable recording medium in which a program for causing a computer to realize the function of quantizing by using a computer is recorded. 22. A computer read recording a program for causing a computer to realize an image forming processing function of quantizing input data of an m-valued multi-gradation image into an n-valued (3 ≦ n <m) by an error diffusion method. A possible recording medium, the function of calculating multi-valued data from the data of the target pixel of the input data and its peripheral pixels, the function of generating a threshold value that changes according to the multi-valued data, and the multi-valued data. Is a predetermined level or higher, a computer-readable recording medium recording a program for causing a computer to realize a function of quantizing using a threshold value so that the occurrence probability of small dots becomes low.