JP4355706B2 - Resolution conversion processing method for binary image - Google Patents

Description

  The present invention relates to a binary image resolution conversion processing method, and more particularly to a binary image resolution conversion processing method for converting a halftone image from a low resolution image to a high resolution image.

  As an example of a conventional binary image resolution conversion technique, Non-Patent Document 1 focuses on the periodicity of image data and uses discrete Fourier transform and discrete inverse Fourier transform, which are frequently used in signal processing, to obtain image data. A method is described in which resolution conversion is performed by converting the period. Specifically, a binary image is converted into a multi-valued image, a discrete Fourier transform process is performed, a surrounding image is cut out from the spectrum image obtained by the discrete Fourier transform, and a value normalization process is performed. Perform discrete Fourier transform processing. Next, the multi-value image data obtained by the inverse discrete Fourier transform process is binarized to create a binary image after resolution conversion. The binarization processing includes fixed value binarization, error diffusion, dithering, and the like.

  On the other hand, in the binary image resolution conversion method described in Patent Document 1, a resolution conversion table is created based on the resolution before and after the linear density conversion of the binary image, and the post-conversion image is obtained from the binary data in each dot of the pre-conversion image. The score is calculated by the ratio of black in each dot. Then, a resolution-converted image is obtained by performing black / white determination after conversion from the calculated black point of each dot and the set determination threshold.

"Gravitational rearrangement of halftone dots for moire-free color proofing" Image Electronics Society of Japan, Vol. 13, No. 5 (2005) JP 2001-358937 A

  However, in the conventional method described in the above-mentioned document etc., since the processing is based on information such as period value conversion and black pixel ratio, the image size after resolution conversion, black pixel density, pixel position, etc. There was a problem that the resolution could not be converted while maintaining it. In addition, the method described in Non-Patent Document 1 has a problem in that a long time is required for processing because the amount of image processing is increased by performing two-dimensional discrete Fourier transform and inverse two-dimensional discrete Fourier transform.

  Therefore, the present invention has been made in view of the problems in the above conventional binary image resolution conversion processing method, and is a binary image resolution conversion method capable of obtaining a high-quality image in a short time. The purpose is to provide.

In order to achieve the above object, the present invention is a binary image resolution conversion processing method, and when converting a binary image from a low resolution image to a high resolution image, the number of pixels after resolution conversion is the same as before resolution conversion. If it is not an integer multiple of the number of pixels of the image, the pixel coordinate calculation using the pixel number ratio of the image before and after resolution conversion and the pixel coordinates of the image before resolution conversion and the adjacent pixels The integer part of the pixel coordinates of the image after resolution conversion corresponding to the coordinates of the pixel to be converted and the part after the decimal point are calculated, and the part after the decimal point is used as a probability value of the amount of change in the number of pixels for expressing the line width. A pixel interpolation process is performed to determine the position of a black pixel or a white pixel of a line image after resolution conversion by randomization correction processing using the probability value .

And according to the present invention, the resolution conversion process is performed for each pixel based on the pixel coordinate information without relying on the image information of the binary image (distribution information of the black and white pixels in a certain area) as in the prior art. It is possible to prevent the processing result from being biased toward black or white pixels and to avoid the occurrence of moire or the like. That is, regardless of the black pixel and the white pixel, the pixels at the same position are subjected to the same resolution conversion process (which is uncorrelated with the distribution of the processing target pixel and the black and white pixels in the region). In addition, moire can be eliminated and the black pixel density can be maintained appropriately even under a situation where the pixel positions of the image before and after resolution conversion are shifted.

In the resolution conversion processing method of the binary image, the interpolation pixel by randomizing the correction processing using the previous SL probability, it is possible to perform each independently in the main scanning direction and the sub-scanning direction. As a result, two-dimensional image processing can be decomposed into two one-dimensional processes, and the amount of computation can be reduced from the conventional x ² to 2x, enabling high-speed conversion processing. As a result, while maintaining the size of the image before the resolution conversion and the image after the resolution conversion, it is possible to suppress the deterioration of the image quality due to the resolution conversion process and perform the high-quality and high-speed binary image resolution conversion.

  In the binary image resolution conversion processing method, the binary image may be a halftone image, and the present invention is used when converting a binary image such as a newspaper image from a low resolution to a high resolution. The processing method concerning can be applied.

  As described above, according to the binary image resolution conversion method of the present invention, a high-quality image can be obtained in a short time when a halftone image or the like is converted from a low-resolution image to a high-resolution image. It becomes.

  Next, embodiments of the present invention will be described with reference to the drawings.

  As shown in FIG. 1, in the present invention, when converting the resolution of a binary image, before converting the resolution, based on the coordinates of the pixel of the image before the resolution conversion and its neighboring (adjacent right and lower) pixels. The pixel position generated in the range of 1 pixel or less in the resolution-converted image by the interpolation process (step S1) of the resolution conversion process for obtaining the pixel of the image after the resolution conversion corresponding to the pixel of the image In order to correct the deviation, randomized correction processing with probability (step S2) for correcting the positional deviation of the pixels independently in the main and sub-scanning directions is performed.

  Next, the overall processing flow of the binary image resolution conversion processing method according to the present invention will be described with reference to FIG.

  First, the dimension information of the image after resolution conversion is obtained from the dimension information of the image before resolution conversion and the resolution of the image after resolution conversion (step S1000), and interpolation processing (step S1100) of resolution conversion processing is performed. After that, randomized correction processing with probability (step S1200) is performed.

  Here, the randomization correction processing with probability (step S1200) includes randomization correction processing with probability in the main scanning direction (step S1210) and randomization correction processing with probability in the sub-scanning direction (step S1220). Further, random correction processing with probability in the main scanning direction (step S1210) includes main scanning direction probability acquisition processing (step S1211), main scanning direction probability generation function processing (step S1212), and main scanning direction randomization correction processing (step S1210). S1213), randomization correction processing with probability in the sub-scanning direction (step S1220), sub-scanning direction probability acquisition processing (step S1221), sub-scanning direction probability generation function processing (step S1222), and sub-scanning direction randomization Correction processing (step S1223).

  Next, the overall operation of the embodiment of the present invention will be described in detail with reference to FIGS.

  First, in step S1000 of FIG. 2 for obtaining the dimension information of the image after resolution conversion, the dimension information of the image before resolution conversion is set to the number of scanning pixels W, the number of sub-scanning lines L, and the resolution R, and after the resolution conversion. Is the number of main scanning pixels W ′, the number of sub-scanning lines L ′, and the resolution R ′, the W, L, R of the image before resolution conversion and the R ′ of the image after resolution conversion. The image information W ′ and L ′ after resolution conversion are calculated from the following equations (1) and (2).

W ′ = W × R ′ ÷ R (rounded down to the nearest decimal point) (1)
L ′ = L × R ′ ÷ R (rounded down to the nearest decimal point) (2)

  Next, interpolation processing (step S1100 in FIG. 2) of resolution conversion processing will be described with reference to FIG.

  First, the pixel coordinates of the image are defined. As for the coordinates of the pixel of the image before the resolution conversion of size W × L, the pixel coordinate at the upper left corner of the image is defined as (0, 0) and the pixel at the lower right corner of the image is defined as (W−1, L−1). . Also, the pixel coordinates of the image after W ′ × L ′ resolution conversion are the same as those of the image before resolution conversion, the pixel coordinates of the upper left corner of the image are (0, 0), and the pixel of the lower right corner of the image is (W '-1, L'-1).

  Next, the pixel coordinates of the corresponding image after resolution conversion are obtained from the pixel coordinates of the image before resolution conversion.

  As shown in FIG. 3, the coordinates s ′ (m ′, n ′) of the image pixel after resolution conversion corresponding to the coordinates s (m, n) of the image pixel before resolution conversion are expressed by the following equation (3): It can be determined using equation (4).

m ′ = m × W ′ ÷ W (rounded down to the nearest decimal point) ・ ・ ・ Equation (3)
n ′ = n × L ′ ÷ L (rounded down to the nearest decimal point) (4)

  Further, resolution conversion interpolation processing using the above equations (3) and (4) will be described.

  In FIG. 3, an image pixel s (m, n) before resolution conversion, a right pixel s0 (m + 1, n) in the vicinity thereof, and a lower pixel s1 (m, n + 1) are expressed by equations (3) and ( 4) is used to obtain pixel coordinates s ′ (m ′, n ′), s0 ′ (m0, n0) and s1 ′ (m1, n1) of the corresponding image after resolution conversion. The image pixel after resolution conversion corresponding to the image pixel s (m, n) before resolution conversion is the pixel s0 ′ (m0, n0) with the image pixel s ′ (m ′, n ′) after resolution conversion at the upper left. ) And s1 ′ (m1, n1) (the pixel in the column with pixel s0 ′ and the pixel in the row with s1 ′ are not included). To the same pixel as the pixel s (m, n) of the image. This process is an interpolation process for resolution conversion.

  Similarly, the image before resolution conversion is raster-scanned, and the resolution-converted image corresponding to each pixel of the image before resolution conversion from pixel (0, 0) to pixel (W-1, L-1) is displayed. Processing for obtaining pixels is performed, and pixel interpolation processing in resolution conversion is performed for all pixels.

  In addition, in order to maintain the image quality (image information) of the image before resolution conversion to the maximum after the resolution conversion, the size of the image after resolution conversion, the density of black pixels, the position information of each pixel, etc. It is necessary to change as little as possible from the image before resolution conversion. However, if the change in resolution of the image before and after resolution conversion does not become an integral multiple, the pixel position of the image before and after conversion is not changed. There is a problem of shifting. This problem causes moiré in the image after resolution conversion.

  For example, when converting an image of 10000 pixels in the main scanning direction with a resolution of 454 dpi to a resolution of 602 dpi, the image after resolution conversion to 602 dpi results in 13200 pixels. Also, if the image before resolution conversion of 454 dpi includes 1000 black pixels independent from each other in 10000 pixels, the image after resolution conversion to 602 dpi maintains the black pixel density. In this case, 1320 pixels out of 13200 pixels must be black pixels. Therefore, the average of 1.32 black pixels for one independent black point of the image before the resolution conversion of 454 dpi is 1.32 black pixels on the average, but in the actual binary image, one black point (by rounding off the decimal point) There are two types of black pixels (matching to integers) and two (matching to integers by rounding up the decimal point) black pixels (the difference in the number of black pixels is one). Further, since the ratio of the two types of black dots is the ratio of the number of pixels of the image before resolution conversion and the number of pixels of the image after resolution conversion, the number of black dots of one black pixel is 68% of the whole. The number of black dots of the two black pixels must be 32% of the total.

  At the black point of one black pixel in the 602 dpi image, the position of the black point is at the center (coordinate 1.0) of the black pixel. Therefore, in the theoretical calculation, the error from the black pixel position of the 454 dpi image is 1.0. −1.32 = −0.32 (unit pixel). That is, as shown in FIG. 4, the pixel position of the image after resolution conversion is shifted to the left by 0.32 pixels from the pixel position of the image before resolution conversion. In addition, in the number of black dots of two black pixels in the 602 dpi image, the black point position is at the center of the two black pixels (coordinate 1.5). Therefore, the theoretical calculation shows an error from the black pixel position of the 454 dpi image. Becomes 1.5−1.32 = + 0.18 (unit pixel). That is, the pixel position of the image after resolution conversion is 0.18 pixels to the right of the pixel position of the image before resolution conversion. From the above, in the case of resolution conversion processing in which the resolution of the image before resolution conversion and the resolution of the image after resolution conversion do not become an integral multiple, the pixel positions of the image before and after conversion are shifted. The problem cannot be avoided.

  Next, as described above, it is used for resolution conversion processing in which the change in the resolution of the image before and after the resolution conversion does not become an integral multiple, in order to solve the problem that the pixel position of the image before and after the resolution conversion is shifted. The randomization correction process with probability (step S1200) in FIG. 2, which is the process, will be described in detail.

  First, a method of expressing pixels after the decimal point in a binary image by adding pixels to randomization will be described with reference to FIG.

  The leftmost end of the figure is a vertical line having a line width of 10 pixels by 10 pixels. At the rightmost end, the line width of 20 pixels is a vertical line of 2 pixels. Between the vertical line with a line width of 1 pixel and the vertical line with a line width of 2 pixels, two black pixels, 5 black pixels, and 8 black pixels are added to the right of the leftmost 1-pixel vertical line. Vertical lines are shown. In terms of density, depending on the number of added black pixels, a vertical line with a source line width of 1 pixel is a vertical line with a line width of 1.2 pixels, a vertical line with a line width of 1.5 pixels, and a vertical line with a line width of 1.8 pixels become. That is, a line width difference of 1 pixel or less that cannot be realized with an actual binary image can be expressed numerically (becomes image calculation) by changing the surrounding pixels to black pixels. By controlling the pixels to be added to randomization for each image pixel after resolution conversion, it is possible to express pixels of 1 pixel or less (pixels after the decimal point) that cannot be originally expressed in a binary image with an apparent density. Become.

  Next, the probability that is a parameter for controlling a pixel to be added to randomization will be described. Taking a vertical line with a line width of 1.2 pixels as an example, a vertical line with a line width of 1 pixel in 10 black pixels is black with a probability of 0.2 for each white pixel of 10 white pixels adjacent to the right. By converting into pixels, 2 pixels out of the total 10 white pixels become black pixels. As a result, in terms of density, the above vertical line having a line width of 1.2 pixels is obtained. Therefore, with the probability that the white pixel adjacent to the black pixel (or the black pixel adjacent to the white pixel) is randomized and converted with reference to the probability (adjusting the appearance probability of the pixel added to the randomization). By randomization correction processing, a line width below an arbitrary decimal point can be expressed by a binary image in theory.

  Next, the randomization correction processing step S1210 with probability in the main scanning direction of FIG. 2 will be described.

  First, the main scanning direction probability acquisition process (step S1211) in FIG. 2 will be described. In this process, the pixel coordinates of the image before the interpolation process of the resolution conversion process corresponding to the pixel coordinates of the image after the interpolation process of the resolution conversion process are estimated, and the estimated pixel coordinates of the image before the interpolation process of the resolution conversion process From this, the amount of deviation between the pixel positions of the image before and after the interpolation processing of the resolution conversion processing is obtained. The amount of pixel position shift obtained here is a random parameter for the randomization correction process.

  For example, the randomization probability k in the main scanning direction of the pixel coordinates (s ′, t ′) of the post-interpolation image in the resolution conversion process is obtained by the following method.

  The pixel coordinate s of the image before the interpolation process of the resolution conversion process in the main scanning direction corresponding to the pixel coordinate (s ′, t ′) of the interpolation process image after the resolution conversion is obtained by the following equation (5).

  s = s ′ × W ÷ W ′ (rounded up to the nearest whole number) (5)

The randomization probability k in the main scanning direction of the pixel coordinates (s ′, t ′) is calculated by the following formula (6) using s.
k = s × W ′ ÷ W (the integer part is rounded down so that only the numerical value after the decimal point is 1 or less) (6)

  Next, the main scanning direction probability generation function processing step S1212 of FIG. 2 will be described.

  First, an integer greater than or equal to 0 is given as an initial value to the seed that is the source of the probability generation function. The seed value indicates the relative position in the main scanning direction. When the image is raster scanned, the value 1 is added each time a pixel is read out.

  In the above step S1211, the obtained probability value k (0 to 1.0) is substituted into the following equation (7) to obtain the occurrence value Q of the probability generation function. Here, the calculation intermediate values I and I 'are integers.

I = seed × (1-k) (rounded down to a whole number)
I ′ = (seed + 1) × (1-k) (rounded down to a whole number) (7)

  Here, when I and I ′ are equal, the probability generation function generation value Q = 1 with respect to the probability value k, and when I and I ′ are different, the probability generation with respect to the probability value k. It is assumed that the function generation value Q = 0. When Q = 1, basically, a pixel is added.

  The generation value of this probability generation function has the following meaning in image processing. For example, when an original image is to be obtained from an image of 100 pixels in which pixels are additionally processed with a probability of 0.15, the number of pixels of the original image is 100 × (1−0.15) = 85. . Since the original image is added as evenly as possible, the pixel coordinate i 'of the added image is 1 to 100, and the corresponding pixel coordinate i of the original image is

i = i ′ × (1−0.15) (rounded down to the nearest decimal point).
In addition, pixels that do not correspond to the pixels of the original image of the added image,
i ′ × (1−0.15) (rounded down) = (i ′ + 1) × (1−0.15) (rounded down)
This is a pixel having coordinates i ′ + 1 that satisfy the above condition, and this is an added pixel (there is no original image pixel corresponding to the pixel i ′ + 1). The added pixels are evenly distributed in the image with a probability of 0.15.

  Further, the randomization correction process with probability in the main scanning direction of FIG. 2 (step S1213) will be described in detail with reference to the flowchart of FIG.

  First, the sub-scanning line position n and the line image data pixel coordinate m of the image after the interpolation process of the resolution conversion process are set to 0th (steps S11 and S12). Next, the image line image after the interpolation processing in the nth row is read (step S13). Further, it is determined whether or not the pixel (m, n) and the right adjacent pixel (m + 1, n) are the same pixel (step S14). When the pixel (m, n) is different from the right adjacent pixel (m + 1, n), the main scanning direction probability acquisition process of the pixel coordinate m is performed (step S15). Randomization correction processing in the main scanning direction is performed on the pixel (m, n) using the acquired probability k (step S16).

  In the randomization correction process in the main scanning direction, the value of the randomization probability k obtained by the equation (6) is substituted into the probability generation function (the above equation (7)), and the generation value Q of the probability generation function is 1. In this case, the pixel (m, n) and the adjacent pixel (m + 1, n) on the right side are converted into the same pixel as the pixel (m, n). When the generation value Q of the probability generation function is 0, pixel conversion processing is not performed. Even when the pixel (m, n) and the right adjacent pixel (m + 1, n) are the same pixel, the pixel conversion process is not performed. Next, 1 is added to the seed value of the probability generation function (step S17).

  Next, it is determined whether or not the line image data is over (step S18). If the line image data has not ended, 1 is added to the pixel coordinate m (points to the next pixel) (step S19). When the line image data ends, the line image data that has been subjected to the conversion process is written in the image after resolution conversion as an image line image after resolution conversion of the nth row. (Step S20). Finally, it is determined whether or not all of the image line images after resolution conversion have been written (step S21). If all of the image line images after resolution conversion have been written, the resolution conversion processing is terminated. If there are remaining line images, 1 is added to the sub-scanning line position n of the image after resolution conversion interpolation processing (refers to the line of the image after interpolation processing for the next resolution conversion) (step S22). This is repeated until writing of the converted image of all the image line images is completed.

  Finally, the randomized correction processing step S1220 with probability in the sub-scanning direction in FIG. 2 will be described.

  First, the sub-scanning direction probability acquisition process (step S1221) in FIG. 2 will be described. In this process, as in the main scanning direction probability acquisition process, the pixel coordinates of the image before the interpolation process of the resolution conversion process corresponding to the pixel coordinates of the image after the interpolation process of the resolution conversion process are estimated, and the estimated resolution conversion From the pixel coordinates of the image before the interpolation process of the process, the shift amount of the pixel position of the image before the interpolation process of the resolution conversion process and after the interpolation is obtained. The obtained shift amount of the pixel position becomes a random parameter for the randomization correction process.

  The randomization probability k ′ in the sub-scanning direction of the pixel (s ′, t ′) of the image after the interpolation processing of the resolution conversion processing is obtained by the following method.

The pixel coordinate t of the image before the interpolation process of the resolution conversion process in the sub-scanning direction corresponding to the image pixel (s ′, t ′) after the interpolation process of the resolution conversion process is obtained by Expression (8).
t = t ′ × L ÷ L ′ (rounded up to the nearest whole number) (8)
Using t, the randomization probability k ′ of the pixel (s ′, t ′) in the sub-scanning direction is calculated by the following equation (9).

k ′ = t × L ′ ÷ L (the integer part is rounded down to a numerical value of 1 or less after the decimal point) (9)

  Next, the sub-scanning direction probability generation function processing step S1222 of FIG. 2 will be described.

  As in the main scanning direction probability generation function processing step S1212 in FIG. 2, first, an integer greater than or equal to 0 is given as the initial value of the seed of the probability generation function. The seed value indicates the relative position in the sub-scanning direction. When the image is raster scanned, the value 1 is added each time a pixel is read.

The probability value k (0 to 1.0) obtained in step S1221 in FIG. 2 is substituted into the following equation (10) to obtain the probability value Q of the probability generation function. Here, the calculation intermediate values I and I ′ are integers.
I (integer) = seed × (1-k) (rounded down to a whole number) (10)

I ′ (integer) = (seed + 1) × (1-k) (rounded down to a whole number)
When I and I ′ are equal, the probability generation function generation value Q = 1 for the probability value k. If I and I ′ are different, the probability generation function generation value Q = 0 is set for the probability value k. When Q = 1, basically, a pixel is added.

  Further, the randomization correction process with probability in the sub-scanning direction in FIG. 2 (step S1223) will be described in detail with reference to the flowchart of randomization correction process with probability in the sub-scanning direction.

  First, the image line image after the interpolation process of the 0th row is read, and the sub-scanning line position n and the line image data pixel coordinate m of the image after the interpolation process of the resolution conversion process are set to 0th (step S31 and step S31). S32). Next, the image line image after the interpolation processing of the (n + 1) th row is read (step S33). Further, it is determined whether or not the pixel (m, n) and the lower adjacent pixel (m, n + 1) are the same pixel (step S34). When the pixel (m, n) and the lower adjacent pixel (m, n + 1) are different, the sub-scanning direction probability acquisition process for the coordinate n is performed (step S35). Using the acquired probability k ', randomization correction processing in the sub-scanning direction is performed on the pixel (m, n) (step S36).

  In the randomizing correction process in the sub-scanning direction, the value of the randomization probability k ′ obtained by the equation (9) is substituted into the above probability generation function (the above equation (10)), and the generation value Q of the probability generation function is In the case of 1, the pixel (m, n) and the lower adjacent pixel (m, n + 1) are converted into the same pixel as the pixel (m, n). When the generation value Q of the probability generation function is 0, the conversion process is not performed. Even when the pixel (m, n) and the lower adjacent pixel (m, n + 1) are the same pixel, the conversion process is not performed.

  Further, 1 is added to the seed value of the probability generation function (step S37). Thereafter, it is determined whether or not the line image data is over (step S38). If the line image data is not the end, 1 is added to the pixel coordinate m (points to the next pixel) (step S39). When the line image data is finished, the line image data that has been subjected to the conversion process is written in the image after resolution conversion as an image line image after resolution conversion of the nth row. (Step S40).

  Finally, it is determined whether or not all of the image line images after resolution conversion have been written (step S41). If all of the image line images after resolution conversion have been written, the resolution conversion processing ends. If there are remaining line images, 1 is added to the sub-scan line position n of the image after resolution conversion interpolation processing (refers to the line of the image after interpolation processing for the next resolution conversion) (step S42). This is repeated until all of the image line images after resolution conversion are written.

As described above, the correction processing of the pixel coordinates of the image after the correction processing of the resolution conversion processing is a randomized correction processing with probability in which the main scanning direction and the sub-scanning direction are separately independent, so that two-dimensional image processing is performed. Can be decomposed into two one-dimensional processes, and the amount of calculation can be reduced from x ² to 2x, enabling high-speed conversion processing.

It is a flowchart which shows the characteristic of the resolution conversion processing method of the binary image concerning this invention. It is a flowchart for demonstrating the whole flow of the resolution conversion processing method of the binary image concerning this invention. It is explanatory drawing of the interpolation process of the resolution conversion processing method of the binary image concerning this invention. FIG. 4 is an explanatory diagram showing that pixel displacement between pre-interpolation and post-interpolation images occurs due to the interpolation processing of the resolution conversion processing of FIG. 3. It is explanatory drawing of the randomization correction process with a probability of the resolution conversion processing method of the binary image concerning this invention. 6 is a flowchart of randomization correction processing with probability in the main scanning direction of the binary image resolution conversion processing method according to the present invention. 6 is a flowchart of randomization correction processing with probability in the sub-scanning direction of the binary image resolution conversion processing method according to the present invention.

Claims (3)

When converting a binary image from a low resolution image to a high resolution image,
If the number of pixels after resolution conversion is not an integer multiple of the number of pixels of the image before resolution conversion, the pixel coordinates before resolution conversion are calculated by pixel coordinate calculation using the ratio of the number of pixels of the image before resolution conversion and after resolution conversion. To calculate the integer part and the decimal part of the pixel coordinates of the image after resolution conversion corresponding to the coordinates of the pixel of the image and the pixel adjacent to the pixel, and to express the line width of the part after the decimal point Conversion to a probability value of the amount of change in the number of pixels of the pixel, and a pixel interpolation process for determining the position of the black pixel or white pixel of the line image after resolution conversion by randomization correction processing using the probability value A resolution conversion processing method for a binary image. The interpolation pixel by randomizing the correction processing using the previous SL probability, resolution conversion processing method of a binary image according to claim 1, characterized in that each independently in the main scanning direction and the sub-scanning direction. The binary image, resolution conversion processing method of a binary image according to claim 1 or 2, characterized in that the halftone image.

Images