JP2014220664A - Image processing apparatus and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique for obtaining a halftone processing result of a high image quality in which deterioration in image quality is reduced, while reducing a jaggy at an edge.SOLUTION: AM screen processing is applied to an input image to generate an image having undergone AM screen processing. When a pixel under consideration in the input image is a pixel detected as a pixel constituting an edge, a value based on the density difference between the input image and the image having undergone AM screen processing in an area including the pixel under consideration is outputted. When the pixel under consideration is a pixel not detected as a pixel constituting the edge, the pixel value of the pixel under consideration in the image having undergone AM screen processing is outputted.

Description

  The present invention relates to a halftone processing technique.

  In general, an image or document created by a PC (personal computer) or the like is printed by a printer. However, the number of gradations of an image that can be printed by a printer is smaller than the number of gradations of an image that can be handled by a PC. For this reason, halftone processing is performed on an image for which printing has been instructed so that the original gradation can be reproduced in a pseudo manner with a certain area where the number of gradations is being reduced.

  By the way, in a laser beam printer which is a kind of printer, an image is formed on a paper surface through complicated processes such as charging, exposure, development, transfer, and fixing. Therefore, for the purpose of securing the gradation by stably fixing the toner on the paper surface, halftone processing may be performed using an AM screen in which threshold matrixes that are concentrated at a certain point where toner dots are formed are arranged periodically. Many. In addition, in order to express color, each color image such as CMYK is subjected to halftone processing using different AM screens, and then each color toner is superimposed to form an image. In such a laser beam printer, in order to reduce as much as possible the color moire that occurs when toners of each color are superimposed, the growth angle of periodically arranged dots called screen angles between AM screens of each color is designed to be as large as possible. It is common to be done.

  In the laser beam printer using the electrophotographic process as described above, when halftone processing is performed using an AM screen, the spatial resolution that can be expressed may be lower than the print resolution. Specifically, an edge portion that is a slanted line in an image may be expressed by dots arranged in a step shape called jaggy. In addition, when the image to be printed includes a periodic pattern such as a halftone dot pattern, an image with interference fringes called document moire is printed due to interference with the periodicity of the AM screen. was there.

  On the other hand, Patent Document 1 discloses a technique for detecting document moire as a density difference by smoothing image data before and after halftone processing and obtaining a difference. Patent Document 2 discloses a technique of performing frequency analysis in the vicinity of a target pixel and determining a mixture ratio of a plurality of halftone processes based on the result. Further, Patent Document 3 discloses a technique in which different halftone processing is performed on data that has been level-corrected so as to enhance or suppress the contour region and data that has been level-corrected for halftone processing, and either one is selected. Yes.

JP 2011-155632 A JP 2010-272987 A JP 2005-341249 A

  With the technique disclosed in Patent Document 1, it is possible to detect document moire caused by interference between a periodic pattern included in image data and the periodicity of the AM screen in the form of a density difference. It is a phenomenon called jaggy that this density difference occurs at the edge portion in the image. Therefore, when this density difference is large, the occurrence of jaggies at the edge portion can be reduced by performing halftone processing that is not an AM screen.

  Further, according to the technique disclosed in Patent Document 2, pixels such as edge portions that require high resolution for expression can be distinguished from each other because they contain a relatively large amount of high-frequency components based on the result of frequency analysis. . Therefore, in a pixel having a large high-frequency component, an attempt is made to express a suitable edge portion by increasing the mixing ratio of the halftone processing result obtained by error diffusion having high resolution as compared with the AM screen.

  Further, in the technique disclosed in Patent Document 3, halftone processing with high resolution is switched for emphasized or suppressed data in the outline portion, and halftone processing with high gradation is switched in the flat portion. By using it, it is trying to express the edge part appropriately.

  However, in these prior arts, since a plurality of different processing results are mixed or switched, pixels generated by mixing or switching may be noticeable and may be visually recognized as image quality degradation. This is because the results of processing with different density expressions are not consistent because mixing and switching are performed without considering each density expression. Further, if mixing or switching is frequently repeated in the image by the above technique, the periodic structure of the AM screen is broken and the screen angle is disturbed, so that color moiré may occur when the color images are superimposed. .

  The present invention has been made in view of such a problem, and provides a technique for obtaining a high-quality halftone processing result with less image quality deterioration while reducing jaggies at an edge portion.

  An image processing apparatus according to the present invention is an image processing apparatus that performs screen processing on an input image, and includes detection means that detects pixels constituting an edge in the input image, and AM screen processing on the input image. Screen processing means for generating an AM screen processed image, and if the target pixel in the input image is a pixel detected as a pixel constituting an edge by the detection means, the input image in the region including the target pixel If a value based on a density difference from the AM screen processed image is output and the target pixel is not detected as a pixel constituting an edge by the detection means, the target in the AM screen processed image is output. Output means for outputting the pixel value of the pixel.

  According to the configuration of the present invention, it is possible to obtain a high-quality halftone processing result with less image quality deterioration while reducing edge jaggy.

FIG. 3 is a block diagram illustrating an example of a functional configuration of the image processing apparatus. The figure explaining an example of the detection technique of an edge. The figure which shows an image group. The figure explaining AM screen processing. The figure which shows the relationship between AM screen process result and a gradation number restoration process result. The figure explaining a reference range. The figure explaining input-output characteristics. FIG. 3 is a block diagram illustrating an example of a functional configuration of the image processing apparatus. The figure which shows an image group. FIG. 3 is a block diagram illustrating an example of a functional configuration of the image processing apparatus. The figure which shows an output image. The figure which shows the hardware constitutions of the apparatus which functions as an image processing apparatus.

  Preferred embodiments of the present invention will be described below with reference to the accompanying drawings. The embodiment described below shows an example when the present invention is specifically implemented, and is one of the specific examples of the configurations described in the claims.

[First Embodiment]
The image processing apparatus according to the present embodiment generates and outputs an output image composed of pixels having pixel values of M (M <N) from an input image composed of pixels having pixel values of N (N ≧ 3). To do. First, a functional configuration example of the image processing apparatus according to this embodiment will be described with reference to the block diagram of FIG.

  The image processing apparatus receives an input image composed of pixels having N (N ≧ 3) pixel values. The edge detection unit 101 detects an edge from the input image that has been input (that is, detects an edge pixel that is a pixel constituting the edge). Techniques for detecting edges are well known, and known techniques such as Laplacian and frequency analysis near the pixel of interest can be used. An example of the edge detection technique will be described with reference to FIG.

  A rectangle marked with “*” in FIG. 2 is a pixel (target pixel) that is a target for determining whether it is an edge pixel or a non-edge pixel. In addition, rectangles in which “a”, “b”, “c”, and “d” are written are pixels adjacent to the target pixel (adjacent pixels). In order to determine whether or not the target pixel is an edge pixel, first, an absolute value of a difference between pixel values of the target pixel and each adjacent pixel is calculated. In FIG. 2, the absolute value of the difference between the pixel values of the adjacent pixel a and the target pixel is represented as | a− * |, and the absolute value of the difference between the pixel values of the adjacent pixel b and the target pixel is represented as | b− * |. Yes. Also, the absolute value of the difference between the pixel values of the adjacent pixel c and the target pixel is represented as | c− * |, and the absolute value of the difference between the pixel values of the adjacent pixel d and the target pixel is represented as | d− * |. Then, the maximum value DIFF is specified from the calculated absolute values of the four differences, and if the specified maximum value DIFF is equal to or greater than the prescribed threshold value THRESHOLD, it is determined that the target pixel is the edge pixel EDGE. On the other hand, if the specified maximum value DIFF is less than the prescribed threshold value THRESHOLD, it is determined that the target pixel is a non-edge pixel non-EDGE.

  With such a technique, it is possible to determine whether each pixel in the image is an edge pixel or a non-edge pixel, and as a result, an edge (edge pixel) can be detected from the image. .

  FIG. 3B shows a result obtained by applying the edge detection technique described with reference to FIG. 2 to the input image when the input image is the image shown in FIG. In FIG. 3B, when the pixel value of only the non-edge pixels among the pixels constituting the input image data is updated to 0, the image (edge image) in FIG. 3B is obtained. In the edge image of FIG. 3B, since the original pixel values remain only for the edge pixels, as a result, an image from which the contour in the input image is extracted is obtained as the edge image.

  Returning to FIG. 1, the AM screen processing unit 102 compares each pixel value in the input image input to the image processing apparatus with each threshold value of the AM screen, and inputs an N-value pixel value. The image is converted into an M-value image. Since this conversion technique is well known, this conversion technique will be briefly described here. Here, an example will be described in which an AM screen that performs ternarization as shown in FIG. 4 is used for the input image (256 gradations) shown in FIG.

  The AM screen shown in FIG. 4 has a structure in which a set of pixels assigned with a set of index numbers called cells is periodically arranged. Each pixel value (input pixel value) of the input image is compared with a threshold value for quantization corresponding to the index number of this AM screen. Depending on the comparison result, 0, 1, or 2 is determined as the output pixel value corresponding to the input pixel value. When AM screen processing is performed on the input image shown in FIG. 3A, an image (AM screen processed image) as shown in FIG. 3C can be obtained. In FIG. 3C, the densities of the output pixel values 0, 1, and 2 in the AM screen processed image are represented as densities corresponding to the input pixel values of 0, 128, and 255. In the image shown in FIG. 3C, the number of pixels included in the cell is 16, and the density other than 0 per pixel is expressed in two levels. Therefore, the first floor of 2 × 16 = 32 gradations and all the pixels of the cell are at the zero level. By adding keys, 33 gradations can be expressed in a pseudo manner in units of cells. However, if the gradation is expressed in units of cells in order to express the number of gradations as much as possible, the resolution is reduced. For this reason, the result of the AM screen processing shown in FIG. 3C is a blurred image due to step-like jaggies and line breaks in the contours and characters in the figure.

  The gradation number restoration unit 103 restores the M-value AM screen processed image converted by the AM screen processing unit 102 to an N-value image. As a result of processing of the AM screen processing unit 102, the density difference detection unit 104 in the subsequent stage detects a density difference from the input pixel value. However, since the number of gradations is reduced by the AM screen processing unit 102, the input pixel value The amount of information is less than the amount of information. For this reason, the number of gradations restoration unit 103 restores the reduced number of gradations to the original number of gradations, thereby restoring the amount of information and improving the calculation accuracy. In the present embodiment, the pixel values 0, 1, and 2 of each pixel in the AM screen processed image may be converted to 0, 128, and 255. To restore the number of gradations, a generally known inverse quantization method may be applied.

  Here, each pixel value in the AM screen processed image is ternary, and the gradation number restoring unit 103 restores each ternary pixel value to a 256 pixel value. FIG. 5 shows a relationship between each pixel value in the AM screen processed image and each pixel value as a result of the gradation number restoration processing.

  FIG. 5A shows the corresponding relationship when the correspondence between the AM screen processing result and the gradation number restoration processing result is determined. However, when the correspondence is determined in advance, the gradation number restoration unit 103 obtains the gradation number restoration processing result from the AM screen processing result based on such correspondence.

  On the other hand, if the correspondence between the AM screen processing result and the gradation number restoration processing result is not known, as shown in FIG. It is also possible to obtain the gradation number restoration processing result by repeating from. However, in this case, a conversion error may be included when the number of gradations is converted. Further, in consideration of the fact that the AM screen processing unit 102 performs quantization in units of cells so that the local density matches the input image, the AM screen processing unit 102 is based on the in-cell average of the processing results of the AM screen processing unit 102. The number of gradations may be restored. Even if the number of gradations is restored based on the average value in the cell, even if the average value in the cell includes a value after the decimal point, by interpolating between the integer values based on the conversion table shown in FIG. The gradation number restoration processing result can be calculated. Hereinafter, the image restored by the tone number restoration unit 103 is referred to as a restored image.

  The density difference detection unit 104 satisfies all 0 ≦ x ≦ X−1 and 0 ≦ y ≦ Y−1 (X and Y are the numbers of pixels in the x direction and the y direction of the image (input image and restored image), respectively). The following processing is performed for x and y. That is, the density difference between the peripheral region of the pixel (x, y) in the restored image and the peripheral region of the pixel (x, y) in the input image is detected. By the AM screen processing performed by the AM screen processing unit 102, a quantization error occurs in most pixels. Therefore, the density difference detection unit 104 calculates the density difference of the AM screen processed image based on the quantization error of each pixel in a local area including the target pixel (hereinafter referred to as a reference area).

  In this embodiment, the reference pixel value is an N-value input pixel value. For this reason, the tone-number restoring unit 103 restored the screen processed image of the AM value to the N value. Further, the AM screen processing unit 102 performs quantization so that the density represented by the input image is represented in cell units. If the input value is not constant in the pixel group constituting the cell, the input image and the AM screen may interfere with each other, resulting in density fluctuation (moire) and image deterioration. Therefore, it is necessary to calculate the density difference that causes the density fluctuation with reference to a local region including the target pixel. Furthermore, when obtaining the density difference, it is preferable to refer to a reference area in a range larger than the cell size on the AM screen. In this embodiment, since AM screen processing is performed in units of 4 × 4 cells shown in FIG. 4, pixels in a range as shown in FIG. 6 are referred to in order to detect density differences.

  In FIG. 6, each rectangle represents a pixel, and the shaded rectangle represents a pixel of interest. In addition, the numbers written in each rectangle are weighting coefficients for calculating the sum of quantization errors in each pixel. When detecting a density difference, a reference region that is as isotropic as possible with respect to the pixel of interest provides a better result without being affected by the periodicity of the input image. For this reason, it is preferable to calculate the sum of quantization errors in reference pixels within the same distance from the target pixel as a density difference. However, in the case of this embodiment, the shape of the cell is a square of 4 pixels × 4 pixels. For this reason, the range shown in FIG. 6A is obtained by determining the reference region so that the pixel of interest is near the center. FIG. 6B shows a region that is larger than a cell of 4 pixels × 4 pixels and isotropic with respect to the pixel of interest. Furthermore, the degree of overlap when the pixel of interest is the 4 × 4 center shown in FIG. 6A is defined as a weighting coefficient. Thus, an area that does not necessarily match the cell shape may be determined as the reference area.

  Here, the density difference in the reference region shown in FIG. 6B is detected. That is, in each of the input image and the restored image, a total of 25 pixels including the pixel of interest (rectangle indicated by diagonal lines) and 24 surrounding pixels (peripheral pixels) are referred to. In the case of FIG. 6B, assuming −2 ≦ i ≦ 2 and −2 ≦ j ≦ 2, first, D1 is obtained by calculating the following equation.

D1 = ΣW (i, j) × S (x + i, y + j)
Here, S (x + i, y + j) is a pixel value in the pixel (x + i, y + j) in the restored image, W (i, j) is a weighting coefficient for the pixel (x + i, y + j), Σ is i = −2 to 2 and This is an operator for calculating the sum for j = −2 to 2. Here, i and j are variables representing pixel positions in the reference area with the target pixel position as the origin. In the case of FIG. 6A, −2 ≦ i ≦ 1 and −2 ≦ j ≦ 1, and in the case of FIG. 6B, −2 ≦ i ≦ 2 and −2 ≦ j ≦ 2.

  However, in the case of FIG. 6B, for example, W (−2, −2) is the weighting coefficient “¼” for the pixel in the upper left corner of the surrounding area, and W (0, 0) is the pixel of interest in the surrounding area. The weighting coefficient “1” for the pixel at the position. Next, D2 is calculated | required by calculating the following formula | equation.

D2 = ΣW (i, j) × T (x + i, y + j)
Here, T (x + i, y + j) is a pixel value at a pixel (x + i, y + j) in the input image. Then, the density difference D (x, y) with respect to the target pixel position (x, y) is obtained by calculating the following expression.

D (x, y) = D1-D2
Here, the difference between W (i, j) × S (x + i, y + j) and W (i, j) × T (x + i, y + j) is obtained by multiplying the quantization error in the pixel (x + i, y + j) by a weighting coefficient. It is a result. That is, the density difference D (x, y) is a value obtained by averaging the quantization error of each pixel included in the reference area with respect to the reference area corresponding to the target pixel. By performing such processing for all x and y satisfying 0 ≦ x ≦ X−1 and 0 ≦ y ≦ Y−1, the density difference D (x, y) for each pixel (x, y) is obtained. Can be sought.

  The density difference adding unit 105 updates the input image by adding the density difference D (x, y) to the pixel value T (x, y) of the pixel (x, y) in the input image.

  The gradation number reducing unit 106 converts the input image (updated input image: N value) updated by the density difference adding unit 105 into an M value image (converted image). The pixel value output from the density difference adding unit 105 is the number of gradations N, and it is necessary to reduce the gradation of this pixel value to the pixel value of M value that is the output of the image processing apparatus. In the present embodiment, an input image having 256 gradations shown in FIG. 3A is converted into ternary values (0, 1, 2) for each pixel using input / output characteristics as shown in FIG. That is, each pixel in the updated input image is referenced, and if the pixel value c of the referenced pixel is 0 ≦ c <64, 0 is output as the output pixel value. If the pixel value c of the referenced pixel is 64 ≦ c <192, 1 is output as the output pixel value. If the pixel value c of the referenced pixel is 192 ≦ c ≦ 255, 2 is output as the output pixel value. In this way, any one of 0, 1, and 2 can be output for all x and y that satisfy 0 ≦ x ≦ X−1 and 0 ≦ y ≦ Y−1. As a result, 256 An update input image of gradation can be converted into an image of three gradations.

  The pixel selection unit 107 sets the output pixel value of each pixel constituting the output image to an image in which the number of gradations is reduced by the gradation number reduction unit 106 and an image in which the number of gradations is reduced by the AM screen processing unit 102. Either of these is selected and output. For this selection, the edge detection result by the edge detection unit 101 is used. That is, if the pixel (x, y) in the input image is an edge pixel, the pixel value of the pixel (x, y) in the image with the number of gradations reduced by the gradation number reduction unit 106 is selected and output. The pixel value of the pixel (x, y) in the image is output. If the pixel (x, y) in the input image is a non-edge pixel, the pixel value of the pixel (x, y) in the AM screen processed image by the AM screen processing unit 102 is selected, and the pixel ( x, y) are output as pixel values.

  Even when the pixel (x, y) in the input image is a non-edge pixel, the density difference may be large. However, when the AM screen has a broken periodic structure, it is easily visible on the flat part. Furthermore, when the AM screen processed images for each color are overlapped to form a color image, there is a concern about image quality deterioration due to occurrence of color moire or the like. Therefore, in the case of a non-edge pixel, the pixel value of the pixel (x, y) obtained by the AM screen processing unit 102 is selected instead of the pixel value updated based on the density difference.

  Further, when the pixel (x, y) in the input image is an edge pixel, the jaggy at the edge can be corrected by selecting the input pixel value corrected based on the density difference.

  If edge detection is performed on the input image shown in FIG. 3A and an input pixel value is selected for an edge pixel and a pixel value in an AM screen processed image is selected for a non-edge pixel, FIG. The output image shown in (d) is obtained. In the output image shown in FIG. 3D, since the input pixel value is selected as the edge pixel, the resolution of the input image is maintained. Further, since the pixel value in the AM screen processed image is selected for the non-edge pixel, the gradation of the input image is maintained. However, for example, in the area A in FIG. 3D, the density appears to be excessively high as compared with the input image. Further, in the region B in FIG. 3D, the density of the edge portion is insufficient and the stepped shape is still conspicuous.

  In this regard, in the present embodiment, an output image as shown in FIG. 3E can be obtained by the above configuration. The output image shown in FIG. 3E has little local density fluctuation as described above. This is because the pixel value is corrected by the density difference that causes moire. Further, in the output image shown in FIG. 3E, the screen structure is maintained while correction using the density difference is performed. The edge is difficult to see even if the screen structure collapses. On the other hand, in the flat part, it becomes noticeable when the screen structure is broken. Therefore, in the present embodiment, the correction based on the density difference is performed only on the edge pixel. For this reason, as a result of the correction, it is possible to suppress the deterioration of the image quality due to the collapse of the screen structure. Since the screen angle of the non-edge portion (flat portion) is not disturbed, color moiré does not stand out even if output images corresponding to the respective colors are superimposed.

  As described above, the AM screen processing result and the result corrected based on the density difference are switched according to the edge detection result. As a result, it is possible to realize halftone processing that maintains the gradation of the flat portion and the resolution of the edge portion and reduces jaggy and moire. Further, by performing correction based on the density difference only on the edge pixels, an output result in which the disturbance of the screen structure is not noticeable can be obtained.

  The configuration described above is merely an example of the basic configuration shown below, and any configuration may be adopted as long as the configuration results in the following basic configuration. That is, the basic configuration image processing apparatus is an image processing apparatus that performs screen processing on an input image.

  In such an image processing apparatus, an AM screen process is performed on an input image to generate an AM screen processed image. For each pixel, the density difference is calculated by averaging the quantization error of each pixel included in the reference area larger than the cell.

  If the pixel of interest in the input image is a pixel constituting an edge, a pixel value obtained based on the density difference is selected. On the other hand, if the pixel of interest in the input image is a pixel constituting a non-edge, a pixel value in the AM screen processed image is selected. Then, the selected pixel value is output as the pixel value of the pixel of interest in the output image.

  In this embodiment, the AM screen processing unit 102 performs ternarization, and the AM screen processed image is converted into image data composed of 0, 1, and 2 pixel values. Then, in order to calculate the quantization error for each pixel, the tone number restoration unit 103 converts 0, 1, and 2 into 0, 128, and 255, respectively, following the AM screen processed image as an input image. However, the AM screen processing unit 102 may be configured to substantially ternarize the input image into the quantization representative values 0, 125, and 255. Thereby, the gradation number restoration unit 103 can be omitted.

  In the present embodiment, an example in which the density difference D (x, y) is calculated by calculation is shown, but a configuration in which filter processing is performed may be used. A filter in which a weighting coefficient is stored in each pixel is used for the reference region for calculating the density difference. Filter processing is performed on each of the input image and the restored image, and the difference between the filter processing results can be acquired as a density component. Alternatively, the same density difference can be calculated by calculating a difference between the input image and the restored image for each pixel, outputting a quantization error image, and performing a filtering process on the quantization error image.

[Second Embodiment]
The image processing apparatus according to the present embodiment detects a density difference by a method that is simpler than that of the first embodiment in image processing. A functional configuration example of the image processing apparatus according to the present embodiment will be described with reference to the block diagram of FIG. In FIG. 8, the same reference numerals are given to the same functional units as those in FIG. In the following description, only differences from the first embodiment will be described, and unless otherwise noted, the same as the first embodiment.

  The gradation number reduction unit 106 according to the present embodiment converts an N gradation input image into an M gradation image. In the present embodiment, an example in which an input image having 256 gradations shown in FIG. 3A is converted into an image having three gradations (an image having ternary values (0, 1, 2) as pixel values) is shown. . When the tone number conversion is performed according to the input / output characteristics as shown in FIG. 7, the input image shown in FIG. 3A obtains an image as a result of the tone number reduction process as shown in FIG. 9A. be able to. The gradation number reduction process result shown in FIG. 9A is similar to the AM screen process result shown in FIG. 3C. The density of the output pixel values 0, 1 and 2 is changed to the input pixel values 0, 128 and 255. Corresponding concentration. In the gradation number reduction processing result shown in FIG. 9A, gradation is not maintained because gradation is performed in units of pixels, and gradation or the like cannot be appropriately expressed. . However, it can be seen that shapes that require high resolution, such as contour portions and characters of the input image, can be faithfully expressed. As described above, the gradation number reduction unit 106 is desirably a gradation number reduction process with high resolution. For example, the gradation number reduction unit 106 may use AM screen processing having a higher line number than the AM screen applied by the AM screen processing unit 102 of the present embodiment.

  The density difference detection unit 104 detects the density difference of the pixel of interest from the AM screen processed image by the AM screen processing unit 102 and the input image. As described above, the density difference of the target pixel is calculated based on the quantization error of each pixel in a local region (reference region) including the target pixel. In this embodiment, in order to calculate the quantization error, the gradation number reduction unit 106 reduces the gradation number of the input image and matches the gradation value of the AM screen processed image. That is, the density difference detection unit 104 performs processing on the image output from the gradation number reduction unit 106 and the AM screen processed image.

  Furthermore, in the present embodiment, the density difference of the target pixel does not include the edge pixel quantization error. The range shown in FIG. 6B is a local region including the target pixel. For the AM screen processed image, the pixel values of the pixels determined as non-edge pixels by the edge detection unit 101 among the pixels included in the reference area are summed. Similarly, for the image with the reduced number of gradations, the pixel values of the non-edge pixels among the pixels included in the reference area are summed. Then, the difference between the summed pixel values is output as the density difference of the pixel of interest.

  The density difference addition unit 105 adds the density difference D (x, y) to the pixel value of the pixel (x, y) in the image whose number of gradations has been reduced by the gradation number reduction unit 106, thereby adding the image. Update. The pixel selection unit 107 selects each pixel constituting the output image from either an image updated by the density difference addition unit 105 (updated image) or an image with the number of gradations reduced by the AM screen processing unit 102. Output. For this selection, the edge detection result by the edge detection unit 101 is used. That is, if the pixel (x, y) in the input image is an edge pixel, the pixel value of the pixel (x, y) in the updated image by the density difference adding unit 105 is selected, and the selected pixel value is output to the output image. The pixel value of the middle pixel (x, y) is output. If the pixel (x, y) in the input image is a non-edge pixel, the AM screen processing unit 102 selects the pixel value of the pixel (x, y) in the image whose number of gradations has been reduced, and the selected The pixel value is output as the pixel value of the pixel (x, y) in the output image.

  Even if the pixel (x, y) in the input image is a non-edge pixel, there is a possibility that a density difference is detected, but in this case, AM screen processing is performed so as not to break the screen structure of the flat portion. The pixel value of the pixel (x, y) in the finished image is selected.

  With the above configuration, when the input image shown in FIG. 3A is processed, the image shown in FIG. 9B can be obtained. The image shown in FIG. 9B is an image in which the jaggies at the edge portions are reduced as compared with the AM screen processed image shown in FIG. Compared with the output image according to the first embodiment shown in FIG. 3 (e), the edge portion becomes an enhanced image. This is because a conversion error is included because the result of the gradation number reduction process is used as the pixel value that serves as a reference when detecting the density difference. On the other hand, since the internal calculation can be performed with an M value having a smaller number of gradations than the N value, the calculation cost can be greatly reduced.

[Third Embodiment]
The image processing apparatus according to the present embodiment detects the density difference with a configuration that is further simplified than in the second embodiment. A functional configuration example of the image processing apparatus according to the present embodiment will be described with reference to the block diagram of FIG. In FIG. 10, the same reference numerals are assigned to the same functional units as those in FIG. In the following description, only differences from the second embodiment will be described, and unless otherwise noted, the same as the second embodiment.

  The density difference detection unit 104 is a first between an image that has been subjected to AM screen processing by the AM screen processing unit 102 and an image in which the number of gradations has been reduced by the gradation number reduction unit 106 that serves as a reference for detecting density differences. The density difference is calculated for each pixel as in the embodiment. That is, with respect to each pixel position, a remaining value obtained by subtracting the total pixel value in the reference in the image whose gradation number has been reduced by the gradation number reduction unit 106 from the total pixel value in the reference area in the AM screen processed image is the pixel. Obtained as the density difference with respect to the position.

  In the present embodiment, the value obtained by converting the input pixel value into the M value by the gradation number reducing unit 106 is converted into the density difference in accordance with the conversion of the N input pixel value into the M value by the AM screen processing unit 102. The calculation cost is reduced by using the pixel value as a reference for detecting the.

  The density difference addition unit 105 determines whether or not to add the density difference to the pixel (x, y) in the input image based on the detection result of the edge detection unit 101. That is, if the pixel (x, y) in the input image is an edge pixel, the density difference adding unit 105 converts the density difference D (x, y) to the pixel value of the pixel (x, y) in the AM screen processed image. ) Is added. The density difference adding unit 105 adds the pixel value obtained by adding the density difference D (x, y) to the pixel value of the pixel (x, y) in the AM screen processed image to the pixel position (x, y) of the output image. Is output as the pixel value at. On the other hand, as a result of this determination, if the pixel (x, y) in the input image is a non-edge pixel, the density difference adding unit 105 determines the pixel value of the pixel (x, y) in the AM screen processed image as The pixel value of the pixel (x, y) of the output image is output.

  By such processing, it is possible to correct pixels generated due to jaggies at the edge portion in the processing result of the AM screen processing unit 102 using the density difference. As a result, an AM screen processing result with reduced jaggy can be obtained. Further, since the screen structure of the flat portion is not broken, it is possible to output an output image in which color moiré is unlikely to occur even when a plurality of color images are superimposed.

  FIG. 11 shows an output image output from the image processing apparatus when the input image shown in FIG. 3A is input to the present image processing apparatus with the above configuration. In the output image shown in FIG. 11, the jaggy of the edge portion is reduced as compared with the AM screen processing result shown in FIG. On the other hand, problems such as line breaks and difficult character reading remain, but this can be improved by performing another process such as extraction of a character region, which is a known technique.

  As described above, in this embodiment, it is possible to reduce jaggy by adding the density difference only to the edge portion of the AM screen processing result. Therefore, it is possible to obtain a halftone processing result with an improved edge portion with a simple configuration without switching between different AM screen processing and halftone processing with higher resolution. In addition, you may use each said embodiment in combination suitably partially.

[Fourth Embodiment]
Each of the functional units shown in FIGS. 1, 8, and 10 may be configured by hardware, but may be implemented as software (computer program). In the latter case, an apparatus capable of executing such a computer program functions as the image processing apparatus according to the first to third embodiments. An example of the configuration of an apparatus capable of executing such a computer program will be described with reference to the block diagram of FIG.

  The CPU 1201 executes operations using computer programs and data stored in the RAM 1202 and the ROM 1203, thereby controlling the operation of the entire apparatus and the image processing apparatus according to the first to third embodiments. The above-described processing is executed.

  The RAM 1202 has an area for temporarily storing computer programs and data loaded from the external storage device 1206 and data received from the outside via an I / F (interface) 1207. The CPU 1201 also has a work area used when executing various processes. That is, the RAM 1202 can provide various areas as appropriate.

  The ROM 1203 stores setting data and a boot program for the apparatus. The operation unit 1204 includes a keyboard, a mouse, and the like, and various instructions can be input to the CPU 1201 by an operator of the apparatus. The display unit 1205 is configured by a CRT, a liquid crystal screen, or the like, and can display the processing result by the CPU 1201 using images, characters, or the like.

  The external storage device 1206 is a large-capacity information storage device represented by a hard disk drive device. The external storage device 1206 is handled as an OS (Operating System), computer programs and data for causing the CPU 1201 to execute the functions of the functional units shown in FIGS. 1, 8, and 10, and known information in the above processing. Information, etc. are stored. In addition, data generated by each of the above processes can be stored. Computer programs and data stored in the external storage device 1206 are appropriately loaded into the RAM 1202 under the control of the CPU 1201 and are processed by the CPU 1201.

  The I / F 1207 functions as an interface for connecting this apparatus to an external device or the Internet. For example, the input image of the N gradations may be input to the apparatus via the I / F 1207, or the output image that is the final output of the apparatus is externally output via the I / F 1207. May be sent.

  Each of the above parts is connected to the bus 1208. The configuration of the apparatus that can execute the computer program for executing the functions of the functional units shown in FIGS. 1, 8, and 10 is not limited to the configuration shown in FIG.

(Other examples)
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, etc.) of the system or apparatus reads the program. It is a process to be executed.

Claims (11)

An image processing apparatus that performs screen processing on an input image,
Detecting means for detecting pixels constituting an edge in the input image;
Screen processing means for performing AM screen processing on the input image to generate an AM screen processed image;
If the pixel of interest in the input image is a pixel detected as a pixel constituting an edge by the detection means, a value based on the density difference between the input image and the AM screen processed image in the region including the pixel of interest And output means for outputting the pixel value of the target pixel in the AM screen processed image if the target pixel is not detected as a pixel constituting an edge by the detection means. A featured image processing apparatus.   The image processing apparatus according to claim 1, wherein the density difference is a value obtained by weighted averaging of quantization errors in pixels included in the region. The output means includes
The density difference is obtained based on a difference between a total pixel value of pixels included in the area in the input image and a total pixel value of pixels included in the area in the AM screen processed image. The image processing apparatus according to claim 2. The output means includes
The value which corrected the pixel value of the said pixel of interest in the said input image about the pixel which comprises the said edge using the said density | concentration difference is output. Image processing apparatus. The output means includes
5. A value obtained by correcting the pixel value of the pixel of interest in the AM screen processed image is output using the density difference for the pixels constituting the edge. An image processing apparatus according to 1. The output means includes
5. The density difference is calculated so as not to include a quantization error of a pixel detected as a pixel constituting an edge by the detection unit among pixels included in the region. An image processing apparatus according to claim 1. The output means includes
Restoring means for generating a restored image by converting the AM screen processed image into an N-value image;
The image processing apparatus according to claim 3, wherein the density difference is calculated based on the restored image and the input image. The output means includes
Conversion means for generating a converted image by converting the input image into an M-value image;
The image processing apparatus according to claim 3, wherein the density difference is calculated based on the converted image and the AM screen processed image.   The image processing apparatus according to claim 1, further comprising a gradation reduction unit that performs gradation reduction on the image output by the output unit. An image processing method performed by an image processing apparatus that performs screen processing on an input image,
A detection step in which the detection means of the image processing device detects pixels constituting an edge in the input image;
A screen processing step in which the screen processing means of the image processing device performs AM screen processing on the input image to generate an AM screen processed image;
If the output means of the image processing device is a pixel in which the pixel of interest is detected as a pixel constituting an edge in the detection step in the input image, the input image and the AM screen processed in the region including the pixel of interest If the pixel of interest outputs a value based on the density difference from the image and the pixel of interest is not detected as a pixel constituting an edge in the detection step, the pixel value of the pixel of interest in the AM screen processed image is An image processing method comprising: an output step for outputting.   The computer program for functioning a computer as each means of the image processing apparatus of any one of Claims 1 thru | or 9.