JP2005311414A - Apparatus, method and program for image processing and recording medium recording this program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus having less amplification of a quantization noise without using a special table to obtain an output image in which the generation of a unique pattern is suppressed; and to provide an image processing method, an image processing program, and a recording medium recording the program. <P>SOLUTION: In gradation data conversion processing (pre-gradation conversion processing: S20) to be performed before color correction processing, gradation data of an input image is converted into a level value having reduced kinds, thereby giving variation to the gradation data of the input image. Then, in gradation data conversion processing (post-gradation conversion processing: S40) by halftone processing, an unprocessed pixel is selected based on the centroid of a cell to generate a cell. At this time, error data exceeding a threshold are returned to a pixel which is lastly incorporated into the cell. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention provides an image that suppresses the occurrence of a peculiar pattern that can occur in an output image of an image output device by converting the tone data of the input image using a predetermined method in an image output device such as a printer. It relates to processing technology.

  Conventionally, in an image output apparatus such as a laser printer or an ink jet printer, gradation data of an input image is binarized with a predetermined threshold value for each pixel, and a predetermined level value is given to a pixel position to which a predetermined level value is given by binarization. An output image is output as a printed matter by forming dots having a size of 1 and printing on printing paper. Here, the process of converting multi-value gradation data into binarized gradation data is generally referred to as halftone processing.

  For halftone processing, depending on the position of the dot to be formed, a unique pattern that does not exist in the input image may occur in the output image, for example, dots may be connected in a chain. In order to suppress the occurrence of such a unique pattern, for example, Patent Document 1 proposes a technique for reducing dot connections. Japanese Patent Application Laid-Open No. 2004-151867 discloses a technique for searching for a pixel by a predetermined table until a total level value of gradation data reaches a threshold value for a pixel of an input image, and generating a dot at a position near the center of the cell. It is disclosed.

  Similarly, in the image output apparatus, color correction processing is usually performed on the gradation data of the input image in accordance with the output color of the image output apparatus. In this color correction process, it has been proposed to reduce the number of types of color correction tables used for the color correction process in order to omit the color interpolation process and speed up the color correction process. That is, Patent Document 2 discloses a technique for performing gradation data conversion processing, that is, processing for reducing the number of level values of input image gradation data to a predetermined number of level values before color correction processing. Has been. Further, Patent Document 2 discloses a technique for performing halftone processing by a predetermined method thereafter.

JP 11-27528 A JP-A-7-30772

  In the halftone processing in Patent Document 1, in order to suppress the occurrence of singular patterns, there has been proposed a method of searching for pixels constituting a cell using a table in which a predetermined search order is recorded. The existing pixels are often already processed, and as a result, the generated cells are distorted (for example, crescent moon type). In a cell having such a shape, the vicinity of the center of the cell may be located outside the cell. For such cells, if dots are formed near the center of the cell, they may overlap with the dots in the adjacent cells, resulting in biased output dots, resulting in a singular pattern. There was a problem of keeping up.

  Further, in Patent Document 1, in order to prevent a regular pattern generated in an output image due to a series of cells having the same shape, a plurality of tables in which the search order of pixels is recorded are prepared, and the table to be used is defined as a cell. A process of switching the cell shape irregularly by switching sequentially or randomly every time has also been proposed. However, this processing has a problem that the processing load is large because a plurality of tables are switched, and the processing speed becomes slow. For example, if the input image is gradation data having approximately the same density, cells having substantially the same shape may be generated even if the search order is random, resulting in a regular unique pattern. There was also a problem of generating.

  In addition, the gradation data conversion processing technique performed before the color correction processing in Patent Document 2 is for reducing the processing load related to color correction of gradation data and increasing the processing speed as described above. For this reason, after gradation data conversion by halftone processing, it was not particularly considered to suppress the occurrence of unique patterns that finally appear in the output image. Therefore, Patent Document 2 describes that both the gradation data conversion process and the halftone process before color correction are performed using the error diffusion method, but in the error diffusion method, In both processes, since the pixel value difference of the pixel of interest that occurs with quantization is distributed in the same direction, there is a problem that quantization noise is amplified and a singular pattern is generated.

  The present invention solves at least a part of such problems and obtains an output image in which the occurrence of singular patterns is suppressed. Therefore, an image processing apparatus, an image processing method, and a low quantization noise amplification without using a special table, An object is to provide an image processing program or a recording medium on which the program is recorded.

  The image processing apparatus of the present invention that achieves the above object is characterized in that Q (Q <P, P is an integer of 2 or more) from gradation data having at least P (P is an integer of 2 or more) kinds of level values for each pixel. ) First gradation data conversion for converting into gradation data having various level values, and M (M is an integer equal to or greater than 2) from gradation data having N (N is an integer equal to or greater than 2) kinds of level values. An image processing apparatus that performs second gradation data conversion for converting to gradation data having various level values, wherein the sum of level values for each pixel is a threshold value in the second gradation data conversion. A pixel group generation unit that generates a pixel group by selecting pixels having the N types of level values until the above is satisfied, and data provision that provides the M types of level values to the predetermined pixels of the generated pixel group The main point is that the

  According to such an image processing device, for example, in the gradation data conversion process performed before the color correction process, the gradation data of the input image is converted into the level value with reduced types, thereby converting the gradation of the input image. Disperse data. Then, in the gradation data conversion process by the halftone process, for the pixels having the gradation data to which the variation is given, the gradation data is changed until the sum of the level values of the gradation data for each pixel becomes equal to or greater than the threshold value. A pixel group (hereinafter referred to as “cell”) is generated by selecting a pixel not subjected to gradation conversion. Then, M kinds of level values are assigned to the predetermined pixels of the cell in which the total level value of the pixels in the generated cell is equal to or greater than the threshold value. In this way, when selecting pixels, the selection order of the pixels varies even if the selection order is not changed for each cell. Therefore, the shape of the generated cell becomes irregular, and the occurrence of a unique pattern is suppressed. A visually comfortable output image can be obtained.

  The image processing apparatus of the present invention further includes a pixel group centroid determining unit that determines a centroid position of the pixel group generated by the pixel group generating unit when performing the second gradation data conversion, and the pixel group generating unit The unit generates the pixel group by continuing to select the pixel closest to the centroid determined by the pixel group centroid determination unit until the sum of the level values for each pixel is equal to or greater than a threshold, and the data providing unit is The M kinds of level values may be given to pixels located at the center of gravity determined by the pixel group center of gravity determination unit.

  In this way, in the cell generation process, the pixel closest to the center of gravity of the cell is always selected and incorporated into the cell, so the shape of the generated cell is approximately circular with the center of gravity of the cell as the center. Since dots are formed at approximately the center position of the circle, for example, the formed inter-dot distance is substantially uniform, the dot dispersibility is improved, and a comfortable output image can be obtained.

  Here, the pixel group generation unit continues to select the pixel closest to the centroid determined by the pixel group centroid determination unit until the sum of the level values for each pixel is equal to or greater than a threshold, and the level in the selected pixel group When the sum of the values exceeds the threshold value, the excess level value may be returned to the pixel selected last by the pixel group generation unit.

  In this way, since the level value of the portion exceeding the threshold value is returned to the last pixel incorporated in the cell that exceeds the threshold value, the level value of the input image before halftone processing is not lost, and the original image and Since the level value of the entire image remains at substantially the same position, it is possible to form a faithful dot in the input image.

  The image processing apparatus of the present invention may use a systematic dither method, an error diffusion method, or an average error minimum method as a method of gradation data conversion when performing the first gradation data conversion. In this way, for example, when the input image is gradation data representing approximately the same density, it can be increased or decreased by converting the level value of the gradation data of the input image to a predetermined level value. As a result, the gradation data varies, and in the subsequent halftone process, which is the second gradation data conversion, cells having substantially the same shape are generated without changing the pixel selection order for each cell. The probability of being reduced is reduced, and a unique pattern is less likely to occur in the output image.

  Further, the image processing method of the present invention that achieves the above object is characterized in that Q (Q <P, P is 2 or more) from gradation data having at least P (P is an integer of 2 or more) kinds of level values for each pixel. First gradation data conversion for converting into gradation data having various level values, and M (M is two or more) from gradation data having N (N is an integer of two or more) kinds of level values. An integer) image processing method for performing second gradation data conversion for converting to gradation data having various kinds of level values, and performing the second gradation data conversion, the sum of the level values for each pixel A pixel group generation step of selecting a pixel having the N types of level values until a threshold value is equal to or greater than a threshold value to generate a pixel group, and a pixel group centroid determination of determining a centroid position of the pixel group generated in the pixel group generation step And the weight determined in the pixel group center of gravity determination step A data providing step for assigning the M kinds of level values to the pixels located at the pixel group, and the pixel group generation step selects a pixel closest to the center of gravity determined in the pixel group center of gravity determination step as a level value for each pixel. The pixel group is generated by continuing to select until the sum reaches a threshold value or more.

  According to such an image processing method, for example, in the gradation data conversion process performed before the color correction process, the gradation data is converted by converting the gradation data of the input image to the level value obtained by reducing the number of types. Can give variation. Then, a cell is generated by selecting a pixel having gradation data given this variation by a predetermined method by gradation data conversion processing by the subsequent halftone processing, and the center of gravity position of the generated cell is set. Form dots. In this way, even if the selection order is not changed for each cell at the time of pixel selection, the probability that cells having the same shape are continuously generated decreases, so that the occurrence of singular patterns can be suppressed visually. A comfortable output image can be obtained.

  Further, the present invention may be a computer program or a recording medium recording the program. That is, for each pixel, gradation data having at least P (P is an integer of 2 or more) kinds of level values is changed to gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. First gradation data conversion to be converted, and gradation data having N (N is an integer of 2 or more) kinds of level values to gradation data having M (M is an integer of 2 or more) kinds of level values An image processing program for performing the second gradation data conversion, wherein when the second gradation data conversion is performed, the N kinds of level values are obtained until the sum of the level values for each pixel becomes equal to or greater than a threshold value. A pixel group generation process for selecting a pixel to generate a pixel group, a pixel group centroid determination process for determining a centroid position of the pixel group generated by the pixel group generation process, and a centroid determined by the pixel group centroid determination process The M types of labels The pixel group generation process is such that the sum of the level values for each pixel is greater than or equal to a threshold value for the pixel closest to the center of gravity determined in the pixel group center of gravity determination process. It is a program for generating the pixel group by continuing to select until. As a recording medium on which the program is recorded, various computer-readable media such as a flexible disk, a CD-ROM, an IC card, and a punch card can be used.

  Furthermore, the image processing apparatus of the present invention that achieves the above-described object provides Q (Q <P, P is 2 or more) from gradation data having at least P (P is an integer of 2 or more) kinds of level values for each pixel. First gradation data conversion for converting into gradation data having various level values, and M (M is two or more) from gradation data having N (N is an integer of two or more) kinds of level values. An integer) image processing apparatus that performs second gradation data conversion to be converted into gradation data having various level values, and the first gradation selection is performed when the second gradation data conversion is performed. A pixel group generation unit that generates a pixel group by selecting pixels having the N types of level values until the sum of the level values for each pixel is equal to or greater than a threshold value based on the reference point; and the pixel group generation unit Pixel group reference point determination for determining the second reference point of the generated pixel group And a data providing unit that provides the M kinds of level values to the pixels located at the second reference point determined by the pixel group reference point determination unit, and the pixel group generation unit includes the first The reference point is updated, and a pixel is selected based on the first reference point.

  According to such an image processing apparatus, a reference point for selecting a pixel to be incorporated into a cell at the time of cell generation is distinguished from a reference point for determining a position of a dot formed in the generated cell. In this way, for example, a pixel to be incorporated can be selected using a reference point specified by the shape of the cell, and a cell can be generated regardless of the level value of each pixel in the cell. As a result, it is possible to select a pixel to be incorporated by paying attention only to the shape of the cell.

  Here, the image group generation unit may update the first reference point every time the pixel is selected a predetermined number of times. In this case, for example, the generated cell is more likely to be formed in a circular shape than when the first reference point is fixed, and a comfortable output image is obtained in which the distance between dots is kept uniform. Is possible.

  Further, the first reference point is a center position calculated from the position of the selected pixel, and the second reference point is a center of gravity calculated from the pixel position of the pixel group and the gradation pixel data value. It may be a position. In this way, selection of pixels to be incorporated into the pixel group may be calculated based on the position of each pixel, and the processing load related to calculation processing can be reduced compared to the case where the center of gravity is calculated. Further, since the position of the dot to be formed is the center of gravity of the generated cell, it is possible to form a dot faithful to the gradation data of the input image in the cell.

  Alternatively, the first reference point may be a center position calculated from the position of the selected pixel, and the second reference point may be a center position calculated from the pixel position of the pixel group. . In this way, for example, in the process of determining the reference point used for pixel selection and the reference point for forming dots, it is not necessary to perform an operation including the level value of the pixel. The processing efficiency of the entire apparatus can be increased.

  Furthermore, the image processing apparatus of the present invention may use a systematic dither method, an error diffusion method, or an average error minimum method as a method of gradation data conversion when performing the first gradation data conversion. . In this way, for example, when the input image is gradation data having approximately the same density, it can be increased or decreased by converting the level value of the gradation data of the input image into a predetermined level value. As a result, the gradation data varies, and cells having substantially the same shape are generated without changing the pixel selection order for each cell in the halftone process that is the second gradation data conversion performed thereafter. The probability of occurrence of a peculiar pattern is less likely to occur in the output image.

  Further, the image processing method of the present invention that achieves the above object is characterized in that Q (Q <P, P is 2 or more) from gradation data having at least P (P is an integer of 2 or more) kinds of level values for each pixel. First gradation data conversion for converting into gradation data having various level values, and M (M is two or more) from gradation data having N (N is an integer of two or more) kinds of level values. An image processing method for performing second gradation data conversion for converting into gradation data having an integer) type of level value, wherein the first gradation selection is performed when the second gradation data conversion is performed. A pixel group generation step of generating a pixel group by selecting pixels having the N types of level values until the sum of the level values for each pixel is equal to or greater than a threshold value based on the reference point, and the pixel group generation step Pixel group reference point for determining the second reference point of the generated pixel group And a data providing step for assigning the M kinds of level values to the pixels located at the second reference point determined in the pixel group reference point determining step, wherein the pixel group generating step includes One reference point is updated, and a pixel is selected based on the first reference point. In this way, for example, a reference point for selecting a pixel to be incorporated into a cell when generating the cell is distinguished from a reference point for determining the position of a dot formed in the generated cell. In this way, for example, a pixel to be incorporated can be selected using a reference point specified by the shape of the cell, and a cell can be generated regardless of the level value of each pixel in the cell. As a result, it is possible to select a pixel to be incorporated by paying attention only to the shape of the cell.

  Furthermore, the present invention may be a computer program or a recording medium recording the program. That is, for each pixel, gradation data having at least P (P is an integer of 2 or more) kinds of level values is changed to gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. First gradation data conversion to be converted, and gradation data having N (N is an integer of 2 or more) kinds of level values to gradation data having M (M is an integer of 2 or more) kinds of level values An image processing program for performing the second gradation data conversion, wherein when the second gradation data conversion is performed, the level value for each pixel is determined based on the first reference point for pixel selection. A pixel group generation process for generating a pixel group by selecting pixels having the N types of level values until the sum exceeds a threshold value, and a second reference point of the pixel group generated by the pixel group generation process is determined. Pixel group reference point determination processing and pixel group reference point determination And causing the computer to execute a data providing process for assigning the M kinds of level values to the pixels located at the second reference point determined in the process, and the pixel group generation process updates the first reference point. And a program for selecting a pixel based on the first reference point. As a recording medium on which the program is recorded, various computer-readable media such as a flexible disk, a CD-ROM, an IC card, and a punch card can be used.

  Next, the best mode for carrying out the present invention will be described.

  FIG. 1 is a block diagram showing the entire system to which the present invention is applied. As a whole, the host computer 10 and the image output device 20 are configured.

  The host computer 10 includes an application unit 11 and a rasterizing unit 12. The application unit 11 generates output image data to be printed, such as character data, graphic data, and bitmap data. For example, the host computer 10 generates character data, graphic data, and the like by operating a keyboard or the like using an application program such as a word processor or a graphic tool. The generated data is output to the rasterizing unit 12. Of course, the host computer 10 may accept data output from an image input device (not shown) such as a scanner and output this data to the rasterization unit 12 as data generated by the application unit 11.

  The rasterizing unit 12 converts the print target data output from the application unit 11 into 8-bit gradation data for each pixel (or for each dot). In the present embodiment, 8-bit gradation data is output for each pixel for each color of R (red), G (green), and B (blue). Therefore, the output gradation data has 256 kinds of values (level values) from 0 to 255 for each pixel. The gradation data generation processing in the rasterizing unit 12 is actually performed by a driver installed in the host computer 10. The gradation data output from the rasterizing unit 12 is output to the image output device 20.

  The image output apparatus 20 includes an image processing unit 30 and a print engine 40 as a whole. The image processing unit 30 first performs gradation data conversion processing for converting the number of gradations on the gradation data output from the host computer 10, and then performs color correction and color conversion processing, halftone processing, pulse width. Modulation processing is performed, and the processed data is output to the print engine 40. The print engine 40 actually performs printing on a recording medium such as printing paper.

  The image processing unit 30 includes a gradation data conversion unit 31, a color correction and color conversion unit 32, a halftone processing unit 34, and a pulse width modulation unit 35. The gradation data converting unit 31 receives gradation data of 8 bits each of RGB (256 bits in total) having 256 kinds of level values output from the host computer 10 and receives RGB gradation data having less than 256 kinds of level values. Conversion processing to gradation data is performed. By this conversion processing, the gradation data output from the host computer 10 can be varied. In this embodiment, the error diffusion method is used as the gradation data conversion processing here. The specific contents of this process will be described later.

  Next, the color correction and color conversion unit 32 receives the gradation-converted RGB gradation data, and the color correction table so that good color reproduction can be obtained in the printed matter actually printed by the print engine 40. 33 is used to correct RGB gradation data. Thereafter, color correction processing is performed on the corrected RGB gradation data from RGB to CMY. Here, C represents cyan, M represents magenta, and Y represents yellow.

  The halftone processing unit 34 quantizes the CMY gradation data output from the color correction and color conversion unit 32 (multi-value such as binary or quaternary) using a predetermined type of level value as a threshold value. Value) and output the quantized data. In the present embodiment, as the halftone process, a process (Circular Cell method, hereinafter referred to as CC method) is used in which a cell composed of a plurality of pixels is formed using the center of gravity of the pixel group, and a dot is generated at the center of gravity. Specific contents of the CC method will be described later.

  The pulse width modulation unit 35 receives the quantized data output from the halftone processing unit 34, and generates a laser driving pulse having a predetermined pulse width as driving data for the quantized data. Then, this drive data is output to the print engine 40.

  The print engine 40 includes a laser driver 41 and a laser diode (LD) 42. The laser driver 41 generates control data for driving the laser diode 42 from the input drive data and outputs the control data to the laser diode 42. The laser diode 42 is driven based on the control data output from the laser driver 41, and further, a photosensitive drum and a transfer belt (not shown) are driven to actually print data from the host computer 10 on a recording medium such as printing paper. Will be.

  In FIG. 1, the gradation conversion process for reducing the number of types of gradation data level values performed before such color correction processing is also referred to as pre-gradation conversion processing. Further, when gradation data conversion processing such as halftone processing is performed after pre-gradation conversion processing, such gradation conversion processing is also referred to as post-gradation conversion processing with respect to pre-gradation conversion processing. The pre-gradation conversion corresponds to the first gradation data conversion described in the claims of the present invention, and the post-gradation conversion corresponds to the second gradation data conversion described in the claims of the present invention.

  Next, a specific configuration of the image output apparatus 20 will be described with reference to FIG. Here, in the image output device 20 in FIG. 1, the gradation data conversion unit 31, the color correction and color conversion unit 32, the color correction table 33, the halftone processing unit 34, and the pulse width modulation unit 35 are the CPU 22 in FIG. Corresponds to the hard disk 25, ROM 26, and RAM 27.

  As a whole, the image output apparatus 20 includes an input interface (I / F) 21, a CPU 22, a hard disk 25, a ROM 26, a RAM 27, and a print engine 40, which are connected to each other via a bus. The input I / F 21 serves as an interface between the host computer 10 and the image output apparatus 20. The input I / F 21 receives RGB gradation data from the host computer 10 transmitted by a predetermined transmission method, and is converted into data that can be processed by the image output device 20. The RGB gradation data is temporarily stored in the RAM 27.

  The CPU 22 is connected to the input I / F 21, the hard disk 25, the ROM 26, the RAM 27, and the print engine 40 via the bus, reads out a program stored in the hard disk 25 or the ROM 26, and performs gradation data conversion and halftone described later. Various processing such as processing is performed. The program in which these processes are recorded may be stored in advance in the hard disk 25 or the ROM 26, or may be supplied from the outside by a computer-readable recording medium such as a CD-ROM, and is not shown in the figure. The image may be stored by being stored in the hard disk 25 provided in the image output device 20 via the RW drive. Of course, it may be stored by accessing a server or the like that supplies a program via a network means such as the Internet and downloading data.

  The RAM 27 serves as a working memory for each process executed under the control of the CPU 22 and stores various data after the process. Then, drive data for driving the laser driver 41 of the print engine 40 is also stored. A specific configuration of the RAM 27 will be described later.

  The print engine 40 has the same configuration as that of the print engine of FIG. 1, and receives the drive data stored in the RAM 27 under the control of the CPU 22 and performs the above-described print processing.

  Next, before describing specific processing performed by the image output apparatus 20 configured as described above, how the gradation data output from the host computer 10 is converted by each processing in the image processing unit 30. This will be described with reference to FIG. This is a supplementary explanation of the pre-gradation conversion process and the post-gradation conversion process in the present embodiment, which will be described later, and will be described in advance in order to avoid complicated explanation.

  In FIG. 3, first, R, G, B gradation data Ri, Gi, Bi output from the host computer 10 are input to the pre-gradation conversion unit 31. The gradation data of Ri, Gi, Bi is 256 gradations, that is, gradation data having 256 kinds of level values from 0 to 255. In the present embodiment, the pre-gradation conversion unit 31 converts the R, G, B data from 256 gradations to 17 gradations. The purpose of this conversion process is to intentionally vary the input image data as described above. Of course, this is also for speeding up the color correction processing to be performed next.

  The pre-gradation conversion unit 31 has 17 kinds of level values (0, 15, 31, 47, 63, 79, 95, 111, 127, 143, 159, 175, 181, 207, 223) converted by the gradation conversion process. 239, 255), R, G, B gradation data Rk, Gk, Bk are output. The 17 level values of the pre-gradation conversion are substantially equally spaced, but this need not be the case.

  Next, the color correction and color conversion unit 32 performs color correction processing of Rk, Gk, and Bk data using the color correction table 33, and from R, G, B to C (cyan), M (magenta), A color conversion process to Y (yellow) is performed. The color conversion process is usually performed using a predetermined color conversion table (not shown). In this embodiment, the CMY gradation data of the color conversion table used for color conversion is gradation data of 8 bits for each color, and gradation data Cc, Mc, Yc is gradation data having 256 kinds of level values from 0 to 255.

  Next, in the present embodiment, the post gradation conversion unit 34 will be described as performing binarization processing for controlling dot on / off, and the gradation data Cp, Mp and Yp output gradation data binarized (quantized) into two kinds of level values 0 or 255. Finally, the image output apparatus 20 performs printing by controlling dot size and dot on / off according to the output level value.

  Now, the process of converting the gradation data in this embodiment will be described with reference to the flowchart of FIG. When this processing is started, first, in step S10, input processing of image data output from the host computer 10 is performed. Subsequently, in the present embodiment, pre-gradation conversion processing by the error diffusion method is performed in step S20.

  The pre-gradation conversion process by the error diffusion method will be described with reference to FIGS. In the error diffusion method, a pixel of interest is determined for a pixel of an input image arranged in a dot matrix, and the scanning line is vertically aligned while repeating main scanning in the horizontal direction (from the left side of the input image to the right) for this pixel of interest. By executing sub-scanning that moves in the direction (downward from the top of the screen of the input image), the gradation data is converted into gradation data having a predetermined type of level value for all pixels. By scanning the pixel of interest in this way, as shown in FIG. 5, the upper row and the lower row of the pixel of interest are divided into gradation converted regions and unconverted regions. Further, the left column and the right column of the pixel of interest are also divided into a tone converted region and a tone non-converted region. Basically, a difference value between a predetermined type of gradation-converted level value and the original level value before gradation conversion, that is, an error data value by quantization (hereinafter simply referred to as “error data”) is used as a target pixel. By assigning them to the right and below, the error of the gradation data in the entire input image is reduced.

  As shown in FIGS. 6A to 6D, when assigning difference values, error data is distributed by weighting predetermined pixels with respect to the target pixel. In this way, by weighting and dispersing predetermined pixels, it is possible to give intentional variation to the gradation data of the input image.

  The pre-gradation conversion process to 17 gradations in this embodiment will be described with reference to the flowchart in FIG. 7 and the schematic diagram in FIG. Note that the distribution of error data in this embodiment is performed using the weighting method of FIG. The error data generated in the target pixel is distributed in a ratio of 2 to the right adjacent pixel, 1 to the lower pixel, and 1 to the lower right pixel.

  In this process, attention is first focused on the pixel in the p-th row and the q-th column. The gradation data Ri, Gi, Bi of the R, G, B components of the pixel is represented as dataB [p] [q] for the B component, for example. This is converted to 17 gradations for each of the R, G, and B components. As an example, processing for converting the B component to 17 gradations will be described below. Of course, the processing is basically the same for the G and B components. Note that error data generated by the 17th gradation of the pixel in the p row and the q column is represented as err [p] [q].

Similarly, in this process, the level value (pre_B) of the B component after the pre-gradation conversion process is expressed as follows.
pre_B [0] = 0;
pre_B [1] = 15;
pre_B [2] = 31;
pre_B [3] = 47;
...
pre_B [15] = 239;
pre_B [16] = 255;

Further, 16 types of threshold values used for the pre-gradation conversion process are defined as “slsh [i] (i = 0, 1, 2,..., 15)”, and slsh [i] = (pre_B [i] + pre_B [I + 1]) / 2. In this case, specifically, the threshold value is as follows.
slsh [0] = 7;
slsh [1] = 23;
slsh [2] = 39;
...
slsh [15] = 247;

  When the processing of FIG. 7 is started, first, in step S201, it is determined whether or not the level value of data_B of the pixel of interest in the p-th row and the q-th column of the input image data is smaller than the level value of slsh [0]. Process. As a result of the determination, if the level value of data_B of the pixel of interest is small (YES), the level value of pre_B is set to 0 of pre_B [0] (step S202).

  On the other hand, if the level value of data_B is greater than slsh [0] (NO), in step S203, the level value of the target pixel is any of the set threshold values slsh [0] to slsh [15]. The process of determining whether or not it exists between the threshold values is performed. If a level value exists between any threshold values (YES), a process of setting the level value of pre_B to a level value existing between the threshold values is performed (step S204).

  On the other hand, if the level value of data_B does not exist between the set threshold values (NO), in step S205, it is determined whether or not the value is greater than or equal to the threshold slsh [15]. As a result of the determination, if the level value of data_B is equal to or greater than the threshold value slsh [15] (YES), the level value of pre_B is set to a value 255 of pre_B [16] (step S206).

  Basically, data_B is gradation-converted to any one of 17 kinds of level values by the above processing, but if it is not converted to any of them, that is, if the result of determination in step S205 is NO Terminates processing as a data error.

  Next, a process of calculating err [p] [q] from the level value pre_B and data_B [p] [q] of the gradation data set in any one of steps S202, S204, and S206 is performed (step S207). ). In the next step S208, the calculated err [p] [q] is distributed to predetermined pixels in accordance with the weighting shown in FIG.

  The above processing is performed in accordance with the scanning order of predetermined pixels, and it is determined whether or not all p and q have been completed (step S209). If not completed (NO), p and q are updated (step S210). Then, by sequentially converting the gradation data of the pixels, the processing for all the pixels of the input image is finished (YES), and the pre-conversion processing is finished. Of course, the pixel in the first row and the first column of the input image is set as the first scanning pixel, and then the target pixel is updated according to the main scanning and sub-scanning methods described above.

  The process shown in the flowchart of FIG. 7 will be described as an example with reference to the schematic diagram of FIG. Here, it is assumed that the gradation data whose level value of the B component of the input image is 85 is pre-gradation converted. As shown in FIG. 8A, the pixel in the p-th row and the q-th column is set as a target pixel. Hereinafter, the pixel is expressed as (q, p).

  By comparison with the threshold value, (q, p) is converted into gradation data having a level value 79 as shown in FIG. Then, error data 6 (= 85-79) is distributed according to a weighting method, with error data of 3 in (q + 1, p), 1.5 in (q, p + 1), and 1.5 in (q + 1, p + 1). According to the scanning method, (q + 1, p) is set as the next pixel of interest (FIG. 8 (c), shaded portion), and (q + 1, p) is converted into gradation data having a level value of 95 by comparison with the threshold value (FIG. 8). (D)). Then, according to the weighting method, the error data -7 (= 88-95) is -3.5 in (q + 2, p), -1.75 in (q + 1, p + 1), and -1.75 in (q + 2, p + 1). Distribute each data. Next, according to the scanning method, (q + 2, p) is set as a target pixel (FIG. 8 (e), shaded portion), and (q + 2, p) is converted into gradation data having a level value 79 by comparison with a threshold value (FIG. 8). 8 (f)). The error data 2.5 (= 81.5−79) is an error of 1.25 for (q + 3, p), 0.625 for (q + 2, p + 1), and 0.625 for (q + 3, p + 1) according to the weighting method. Distribute each data. In this way, the gradation data conversion process is repeatedly performed for all the pixels of the input image, and the pre-gradation conversion process is performed.

  As described above, the input gradation data is converted into the input gradation data by the pre-gradation conversion, as is apparent from the comparison between the pixels surrounded by the thick frame shown in FIG. 8F and the pixels shown in FIG. Can give variation.

  Next, returning to FIG. 4, color correction and color conversion processing of the gradation data Rk, Gk, Bk after the pre-gradation conversion is performed in the next step S30. In this embodiment, as described above, the color conversion process converts the gradation data into Cc, Mc, and Yc gradation data having 256 gradations. With the variation given by.

  Next, post gradation conversion processing by the CC method is performed in step S40. The specific processing content of post tone conversion by the CC method will be described with reference to the flowchart of FIG. 9 and the schematic diagram of FIG. Here, description will be made assuming that gradation conversion (binarization) is performed on two kinds of level values of 0 and 255. Note that the processing shown in the flowchart of FIG. 9 is performed on the gradation data of each color component of Cc, Mc, and Yc.

  When post tone conversion is performed, first, in step S401, processing for reading an input image is performed. Here, the gradation data of the color component Cc is read as an input image. An example of the read gradation data is shown in FIG. For example, gradation data having a level value of 50 is given to the pixel in the m-th row and the n-th column (hereinafter referred to as (n, m)).

  Next, in step S402, a process for searching for a pixel for which post-gradation conversion has not been performed is performed, and a process for determining whether there is a search unprocessed pixel is performed. In the search for unprocessed pixels, as in FIG. 5, the search scan is always performed in the horizontal direction first, and then the search scan is performed in the vertical direction. That is, when there is an unprocessed pixel, the leftmost pixel is searched for among the uppermost pixels in the entire input image. Of course, at the start of the post-gradation conversion process, the pixel (1, 1) in the first row and the first column is searched.

  Next, when there is an unprocessed pixel as a result of the determination (YES), the process proceeds to the next step S404. On the other hand, if there is no unprocessed pixel as a result of the determination (NO), the entire input image is post-gradation converted, and a process of outputting a binarized image is performed (step S403), and the post-gradation conversion process is performed. End the routine.

  In step S404, a process for determining whether or not a cell expansion end condition is satisfied is performed. If the result of determination is NO, processing for enlarging the cell is performed (step S405). In the present embodiment, the condition for expanding the cell is that the total level value of the pixels in the cell is less than 255. Further, the cell is expanded by incorporating the pixel closest to the center of gravity of the cell into the cell. On the other hand, if the result of determination is YES, cell expansion is terminated, and processing proceeds to step S406.

  The processing in step S404 and step S405 will be specifically described with reference to FIG. Now, it is assumed that the searched unprocessed pixel is the pixel (n, m) in the m-th row and the n-th column indicated by the shaded portion in FIG. 10A, and all other pixels are also unprocessed pixels. At this time, a cell CL of (n, m) 1 pixel is generated. The position of the center of gravity GC of the cell CL is the center of the pixel (n, m).

  Next, since the sum of the level values of the pixels in the cell CL is 50 and less than 255, the centroid position of the cell CL is calculated, and as shown in FIG. 10B, the calculated centroid GC The unprocessed pixel closest to the position is incorporated into the cell CL. In this case, two of (n + 1, m) (n, m + 1) exist as the pixels closest to the center of gravity, but in this embodiment, (n + 1, m) is incorporated in the cell according to the order in the scanning direction, and the cell CL is enlarged. .

  Next, since the sum of the level values of the pixels in the enlarged cell CL (the portion in the thick frame in FIG. 10C) is 100 and less than 255, the cell position is determined from the pixel position in the cell and the pixel level value. The position of the center of gravity GC of CL is calculated (FIG. 10 (c)). 10C, among the pixels (n, m + 1) (n + 1, m + 1) closest to the calculated position of the center of gravity GC, (n, m + 1) according to the same order in the scanning direction. Is incorporated into the cell, and the cell CL is enlarged.

  Next, the sum of the level values of the pixels in the enlarged cell CL (the portion within the thick frame in FIG. 10D) is 150, which is less than 255, and therefore, from the pixel position in the cell CL and the level value of the pixel. The position of the center of gravity GC of the cell CL is calculated. Then, as shown by the shaded portion in FIG. 10D, the pixel (n + 1, m + 1) closest to the calculated position of the center of gravity GC is incorporated into the cell, and the cell CL is enlarged.

  Next, the sum of the level values of the pixels in the enlarged cell CL (the portion within the thick frame in FIG. 10E) is 205, which is less than 255, and therefore, from the pixel position in the cell CL and the level value of the pixel. The position of the center of gravity GC of the cell CL is calculated. Then, as shown in FIG. 10E, the pixel (n + 1, m + 2) closest to the calculated position of the center of gravity GC is incorporated into the cell, and the cell CL is enlarged.

  Next, the sum of the level values of the pixels in the enlarged cell CL (the portion within the thick frame in FIG. 10 (f)) is 255, and the cell enlargement end condition is satisfied.

  Next, in step S406 (FIG. 9), processing for determining whether or not the total level value of the pixels in the cell is equal to the threshold value is performed. If it is equal to the threshold value as a result of the determination (YES), a process of calculating the position of the center of gravity GC of the cell from the pixel position in the cell and the level value of the pixel is performed (step S407). Then, a process of giving the same level value as the threshold value to the pixel existing at the calculated center of gravity position is performed (step S408), and the process returns to step S402 to repeat the search process for unprocessed pixels.

  As shown in FIG. 10F, in step S408, the level value “255” is given to the pixel (n + 1, m + 1) existing at the position of the center of gravity GC calculated in step S407. By subsequent processing (step S50 (FIG. 4 described later)), a dot having a predetermined color and size is formed at the pixel position to which this level value “255” is given, and an image is printed.

  On the other hand, if the result of determination in step S406 is not equal to the threshold value, that is, if the threshold value is exceeded (NO), processing for calculating the center of gravity of the cell is performed ignoring the remaining error data (step S409). Then, this error data is returned to the pixel that was last incorporated in the cell (step S410), and a process for setting the pixel for which the error data has been returned as an unprocessed pixel is performed (step S411). Then, the process proceeds to step S408, and a process of giving the same level value as the threshold value to the pixel existing at the calculated barycentric position is performed.

  The processing from step S409 to S411 will be specifically described with reference to FIG. In FIG. 11A, following the generation processing of the cell CL1 described in FIG. 10F, in step S402, the pixel (n + 2, m) is searched as an unprocessed pixel according to the scanning direction by a predetermined scanning method. Shows the state. Then, processing is performed in the same manner as described with reference to FIG. 10, the pixel (n + 2, m + 1) in FIG. 11B, the pixel (n + 3, m + 1) in FIG. 11C, and the pixel in FIG. The cell is expanded by incorporating (n + 3, m + 2) into the cell CL2.

  As shown in FIG. 11E, when the pixel (n + 3, m + 2) is incorporated into the cell CL2, the total level value of the pixels in the cell CL2 is 265, which exceeds the threshold 255. Therefore, the excess error data 10 (= 265-255) is ignored, and the center of gravity GC2 of the cell CL2 is calculated assuming that the pixel (n + 3, m + 2) has the level value 50 (= 60-10).

  Then, as shown in FIG. 11F, the level value “255” is assigned to the pixel (n + 3, m + 1) existing at the calculated position of the center of gravity GC2. At this time, the error data “10” is returned to the pixel (n + 3, m + 2) last incorporated in the cell CL2. Then, the pixel (n + 3, m + 2) is set as an unprocessed pixel having a level value of 10, and is excluded from the cell CL2.

  The gradation conversion process based on the CC method described with reference to FIGS. 9, 10, and 11 is performed for all gradation data Cc, Mc, and Yc, and step S <b> 40 is terminated, and the generated “0” and “255” are generated. Quantized data having a level value of is output. In the next step S50 (FIG. 4), a process of outputting as image data for printing is performed.

  In step S50, a laser drive pulse having a predetermined drive pulse width is generated as drive data in order to form a dot at the pixel position to which the quantized data of the level value “255” is assigned. Then, by outputting this drive data to the print engine 40, dots of a predetermined color and size are formed on a printing paper or the like and output as a printed matter.

  The image processing content in the present embodiment has been described above. However, according to the post-gradation conversion processing by the CC method, a cell is always generated by incorporating the pixel closest to the center of gravity of the cell. On the other hand, a cell having a substantially circular shape is generated. Therefore, when a dot is formed at a pixel position to which a threshold value is assigned, the probability that adjacent dots are close or overlapped can be reduced, and the occurrence of a unique pattern due to such overlapping of dots can be suppressed. .

  Further, as described above, in the gradation conversion processing in Patent Documents 1 and 2, error data is transferred to adjacent pixels and cells, so that the input image data is distorted, resulting in a unique pattern that does not exist in the input image. However, according to the post-gradation conversion by the CC method, the error data is returned to the original pixel, so that the distortion of the input image data is reduced, and the gradation data of the input image is reduced. The gradation conversion process can be performed relatively faithfully.

  Furthermore, as in this embodiment, the occurrence of a regular singular pattern can be suppressed by combining the pre-gradation conversion process using the error diffusion method and the post-gradation conversion process using the CC method. . The reason for this will be described next.

  FIG. 12 shows, as an example, the gradation data of the Cc component before the post gradation conversion process. FIGS. 12A to 12C are the gradation data of the input image when the pre gradation conversion process is not performed. (D) to (f) illustrate the gradation data of the input image when the pre-gradation conversion process is performed. Note that some of the level values of the pixels that are not necessary for the description in the gradation data are omitted and left blank for the sake of simplicity.

  When the pre-gradation conversion process is not performed, the color correction and the color conversion process usually only perform the color correction and the color conversion process on the input image data, and the level value of the input image data varies. Therefore, when the input image is gradation data having a uniform density, the gradation data before the post gradation conversion process is inputted with a level value having a substantially uniform density. Therefore, as an example, it is assumed that the gradation data Cc of the color component C (cyan) is gradation data having a uniform level value 85 obtained by simply color-converting the gradation data of the input image. Then, as shown in FIG. 12A, it is assumed that the cell CL3 is generated by being input as gradation data before post gradation conversion.

  Next, a process of selecting a pixel to be incorporated into the cell CL3 is performed. When the pre-gradation conversion is not performed, as shown in FIG. 12B, the center of gravity GC3 has the same level value of the pixels (n, m) and (n + 1, m). Will be located. At this time, there are two pixels (n, m + 1) and (n + 1, m + 1) as pixels to be incorporated into the cell. Then, as shown in FIG. 12 (c), (n, m + 1) is selected according to the scanning direction and incorporated into the cell.

  As described above, when gradation data having uniform level values is subjected to gradation conversion, there are often a plurality of pixels to be incorporated into a cell, and the pixels selected at that time are selected according to a predetermined scanning order. The probability that pixels having substantially the same positional relationship with respect to the cell are selected each time is increased. As a result, cells having the same shape are generated continuously, and the above-described regular unique pattern is likely to occur.

  On the other hand, when pre-gradation conversion processing is performed, color correction and color conversion processing are performed with variations provided, so even if the input image has a uniform density, gradation data before post-gradation conversion processing is used. Is inputted with a level value given a variation. For example, when the gradation data of FIG. 12A is actually pre-gradation converted by the error diffusion method as described in FIG. 8, the gradation data before the post gradation conversion is shown in FIG. It becomes a state. At this time, it is assumed that the cell CL3 is generated.

  Next, in the process of selecting a pixel to be incorporated into the cell CL3, as shown in FIG. 12E, a difference occurs in the level values of the pixels (n, m) and (n + 1, m) by the pre-gradation conversion process. Therefore, the center of gravity GC3 is not in the middle of the pixel, but is located in the pixel (n + 1, m). At this time, only (n + 1, m + 1) pixels are to be incorporated into the cell. As a result, as shown in FIG. 12F, a cell having a shape different from that in FIG. 12C is generated.

  As is clear from the explanation in FIG. 12, when the gradation data of the level value to which the variation is given is subjected to the gradation conversion, the probability that the position of the center of gravity is located in the middle of the pixel due to the variation is reduced, The probability of being located in a pixel having a large level value increases. For this reason, the pixels to be incorporated into the cell will vary appropriately according to the variation in level value. As a result, the probability that cells having the same shape are continuously generated is reduced, and it becomes easy to suppress the occurrence of the regular singular pattern described above.

  Further, as described above, in the pre-gradation conversion, when the level value varies, error data of the level value is distributed to adjacent pixels. Therefore, even in post-gradation conversion, if error data is transferred to adjacent pixels or cells as in Patent Document 1 or 2, the error data for the gradation data of the input image is further increased. In some cases, the quantization noise due to the tone conversion process is amplified to easily generate a singular pattern or the like. As in the present embodiment, by performing the post-gradation conversion process by the CC method, the gradation data of the input image to which the variation is given is scaled without further dispersing the error data generated by the pre-gradation conversion process. Since tone conversion is possible, it is possible to obtain a comfortable output image that is faithful to the gradation data of the input image while suppressing the occurrence of regular patterns.

  Next, in the hardware configuration of the image output apparatus 20 to which this embodiment is applied, the specific processing contents of each hardware in the post-gradation conversion will be described with reference to the flowchart of FIG. 9 and the configuration diagram of FIG. Note that the error diffusion method, color correction, and color conversion processing are known as well-known processing techniques according to Patent Document 1 and the like, and thus description of these processing contents is omitted.

  First, the CPU 22 reads a program stored in the hard disk 25 or the ROM 26, and starts an input image reading process (step S401). The input image is read by fetching the input image data from the host computer 50 into the RAM 27 via the input I / F 21. At this time, input image data of each color component of Cc, Mc, and Yc is stored in an input buffer area that is a predetermined area of the RAM 27. An example of this is shown in FIG.

  As shown in FIG. 13A, the RAM 27 includes an input buffer area 271, a working memory area 272, and an output buffer area 273. The input buffer area 271 has a two-dimensional structure. For example, for the pixel of the input image (hereinafter, input pixel), the input pixel located at (n, m) is (n, m) of the input buffer area. m). An input buffer area 271 is configured to correspond to the position of the input pixel. Actually, the level value of the input image data is stored at this address. For example, when the gradation data of the Cc component input image shown in FIG. 10A is input, the level value of each pixel is stored in the input buffer 271 as shown in FIG. The working memory area 272 stores the calculation result of the center of gravity position and the calculation result of the level value, and the output buffer area 273 is configured to correspond to each input pixel similarly to the input buffer 271. Details will be described later.

  Next, the CPU 22 searches for and determines an initial pixel that is the first unprocessed pixel for cell generation (step S402), and determines whether or not the level value of the determined initial pixel is smaller than a threshold value (step S404). Specifically, the threshold value stored in the ROM 26 is read and compared with the gradation value of the initial pixel read from the input buffer 271 of the RAM 27. If the threshold value is smaller than the threshold value, the process proceeds to step S405. The process proceeds to step S406.

  The initial pixel is determined by the CPU 22 selecting the pixel located on the upper left side in the input pixel position from the input image data stored in the input buffer 271 of the RAM 27. That is, in the example shown in FIG. 13A, the CPU 22 determines that the pixel located at the position corresponding to (n, m) is the initial pixel. Hereinafter, in order to simplify the description, (n, m) = (0, 0) will be described.

  If it is determined in step S404 that the value is smaller than the threshold value (YES), the CPU 22 selects an unprocessed pixel and performs a process of enlarging the cell (step S405). Here, selection of unprocessed pixels is performed using the center of gravity of the cell. By selecting pixels using the center of gravity, the final cell shape is circular, and by forming dots at the center of gravity, the inter-dot distance is uniform and visually comfortable. A printed output can be obtained.

  The CPU 22 reads the level value of the pixel where the initial pixel is located and its address position, and calculates the center of gravity. Specifically, the CPU 22 calculates the center of gravity by using the following arithmetic expression.

(Formula 1)
x _centroid = {{(x coordinate of centroid of cell) × (total level value of pixels in cell)} + {(x coordinate of selected unprocessed pixel) × (level value of selected unprocessed pixel)}} / {(Total level value of pixels in cell) + (level value of selected unprocessed pixel)}
y _centroid = {{(y coordinate of cell centroid) × (total level value of pixels in cell)} + {(y coordinate of selected unprocessed pixel) × (level value of selected unprocessed pixel)}} / {(Total level value of pixels in cell) + (level value of selected unprocessed pixel)}
(X _{center of gravity} , y _center of gravity are coordinates of the _center of gravity)
This arithmetic expression is stored in the ROM 26, and the CPU 22 reads and performs the arithmetic operation in step S405. Of course, in the case of the initial pixel, the x and y coordinates of the center of gravity of the cell are calculated as 0. The CPU 22 stores the calculated barycentric position and the level value 50 of the initial pixel in the above-described working memory area 272 (see FIG. 13B). This is because the subsequent arithmetic processing is performed smoothly.

  If it is determined in step S404 that the level value of the initial pixel is smaller than the threshold value, the CPU 22 stores “0” in the output buffer 273 corresponding to the initial pixel, and the level value of the initial pixel is set to the threshold value in step S404. If it is determined as described above, “255” is stored in the output buffer 273 corresponding to the initial pixel at that time. This is because a process of placing dots on this pixel is performed. Here, as shown in FIG. 13B, “0” is stored by the CPU 22 in the output buffer 273 of the RAM 27 corresponding to the selected pixel.

  The unprocessed pixel selection process is performed by selecting a pixel that is closest to the center of gravity calculated in step S405. At this time, the image data at the address (1, 0) or the image data at the address (0, 1) is selected from the input image data stored at a predetermined position in the input buffer of the RAM 27. This is because the two pixels are equidistant from the address (0, 0). Here, one of the pixels needs to be selected, but in this embodiment, the CPU 22 selects the pixel data at the address (1, 0) in the 0th row and the first column in accordance with the scanning direction. Thus, the process of enlarging the cell is performed.

  Next, the CPU 22 determines again whether or not the total level value of the pixels in the cell is less than the threshold value (step S404). That is, it is determined whether or not the total level value of the pixels in the cell including the level value of the selected unprocessed pixel is less than the threshold value “255”. Specifically, the CPU 22 reads the level value of the pixel written in the working memory 272 of the RAM 27 and the level value stored in the input buffer 271 of the pixel selected in step S405, calculates the sum, and calculates the ROM 26 Is compared with the threshold value stored here (the value of “255” in this case). For example, in the example of FIG. 13C, since the level value stored at the address (1, 0) is “50”, the total level value is “100”. Therefore, the total value does not reach the threshold value.

  When the sum of the level values does not exceed the threshold value (step S404: YES), the CPU 22 calculates the center of gravity including the unprocessed pixel selected again (step S405). The specific calculation of the CPU 22 is performed using the above-described calculation formula (Formula 1). The x and y coordinates of the center of gravity obtained so far are (0, 0), the level value of the center of gravity is “50”, the x and y coordinates of the unprocessed pixels are (1, 0), and the gradation value is “50”. ", The calculated center-of-gravity position is (0.5, 0). The barycentric position calculated in step S405 is stored in the working memory 272 of the RAM 27 together with the sum of the cell level values under the control of the CPU 22. In the previous processing in step S405, the centroid position of the initial pixel and the level value of the initial pixel were stored in the working memory 272, but here the position is overwritten and stored. An example after the storage is shown in FIG.

  Here, the position of the center of gravity is not necessarily an integer value as shown in the above (Equation 1). In addition to selecting the pixel closest to the center of gravity, the pixel closest to the pixel where the center of gravity is located may be selected. By doing so, the position of the center of gravity can be calculated as an integer value, and a reduction in the amount of calculation processed by the CPU 22 can be expected.

  Here, the CPU 22 stores “−1” in the input buffer 271 corresponding to the selected initial pixel. This is because the pixel is not selected again by selecting the unprocessed pixel. Thereafter, in the unprocessed pixel selection process, a process of storing “−1” in the input buffer 271 corresponding to the already selected pixel is performed simultaneously with the selection process or before the selection process. The reason why such processing is performed is that the level value stored in the input buffer 271 is stored in the working memory 272 as the total value and is not required in the subsequent processing.

  The CPU 22 stores “−1” in the input buffer 271. However, the present invention is not limited to this, and other negative values may be used. Further, a memory area having the same configuration as that of the input buffer 271 may be prepared, and a flag may be set at a position corresponding to the selected pixel to determine whether or not the selected pixel has been processed.

  In this way, each time the processing of steps S404 and S405 is repeated, the CPU 22 configures the cell until the threshold value is reached while determining whether or not the total level value of the pixels in the cell calculated in step S404 is equal to the threshold value. Can continue. If the threshold value has not been reached, as described above, the CPU 22 stores “0” in the position of the output buffer 273 corresponding to the position of the input pixel incorporated in the cell.

  FIG. 14A shows a state when the unprocessed pixel (1, 1) is incorporated in the cell, and the CPU 22 reads out the total level value of the pixel in the cell and the threshold value stored in the ROM 26. Then, since the threshold value has not been reached yet, the pixel (1, 1) is incorporated into the cell, and the working memory 272 stores the centroid position and the total level value of the cell as well as the selected unprocessed pixel (1, “0” is stored in the output buffer 273 corresponding to 1).

  FIG. 14B shows a process performed subsequently to FIG. 14A, and shows a state where the input pixel (1, 2) is selected as the pixel closest to the center of gravity position. At this time, the state is shown in which the total level value of the cell is “255”. That is, since the total level value “205” of the pixels incorporated so far in the cell is stored in the working memory 272 of the RAM 27, the total value of the value and the value of the pixel (1, 2) is compared with the threshold value. Thus, the CPU 22 determines that it is equal to the threshold value. In this case, the CPU 22 stores “0” in the position of the output buffer 273 corresponding to the position of the input pixel (1, 2) incorporated in the cell.

  If the CPU 22 determines that the threshold value is equal to the threshold value (step S406: YES), the process proceeds to step S407, the center of gravity position is calculated using (Equation 1), and the calculated center of gravity position and the sum of the cell level values are calculated. Store in the memory 272. Then, a process for forming a dot of a predetermined size on a pixel located at the center of gravity is performed. Specifically, as shown in FIG. 14B, “255” is written at the address position corresponding to the position of the input pixel of the output buffer 273 of the RAM 27. The structure of the output buffer is also a two-dimensional structure like the input buffer, and has an address space corresponding to the position of the input pixel. By writing “255” at the address where the center of gravity is located, it is possible to subsequently hit a dot at a position corresponding to that position.

  Next, the CPU 22 proceeds to a process of generating a new cell by using the writing process of the data “255” as a trigger, and determines again whether there is an unprocessed pixel (step S402). That is, it is determined whether or not the above-described processing has been performed for all level values written in the input buffer 271 of the RAM 27. If there is an unprocessed pixel (YES), the processing proceeds to step S404 again, and the above-described processing is performed. Repeat the process. For example, FIG. 14C shows a state in which the unprocessed pixel (2, 0) is searched for in accordance with the scanning method described above. If the processing is completed for all the pixels (step S402: NO), the CPU 22 proceeds to step S403 and the processing is terminated.

  On the other hand, if the total level value of the pixels in the cell exceeds the threshold value in step S406 (NO), the process proceeds to step S409, and the center of gravity calculation process is performed when the threshold value is exceeded.

  FIG. 15A shows a state in which the processing of the CPU 22 is performed following the processing described in FIG. 14, and shows the RAM 27 when the input pixel (3, 1) is incorporated in the cell. ing. Next, as shown in FIG. 15B, when the pixel located at (3, 2) is selected, the total level value of the pixels in the cell is calculated to be “265”, and the threshold value “255”. Will be exceeded. Therefore, in such a case (step S406: NO), the process proceeds to step S409.

  In step S409, the center value of the cell is calculated by ignoring the level value exceeding the threshold, that is, the error data. Therefore, first, the CPU 22 performs an operation of subtracting the total level value of the pixels in the cell from the threshold value, that is, the level value of the working memory stored in step S405 so far from the threshold value. For example, in the example shown in FIG. 15, the threshold value is “255”, and the level value stored so far in step S405 is “205” as shown in FIG. 15A. = 50 values are calculated.

  Then, the CPU 22 calculates the center of gravity using the value of “50” as the gradation value of the pixel (3, 2), so that the error data is ignored. Specifically, the following calculation formula is used for the calculation of the center of gravity.

(Formula 2)
x _centroid = {{(x coordinate of centroid of cell) × (total level value of pixels in cell)} + {(x coordinate of selected unprocessed pixel) × (calculated value of step S409)}} / {( Sum of level values of pixels in cell) + (calculated value of step S409)}
y- _centroid = {{(y-coordinate of cell centroid) × (total level value of pixels in cell)} + {(y-coordinate of selected unprocessed pixel) × (calculated value in step S409)}} / {( Sum of level values of pixels in cell) + (calculated value of step S409)}
(X _{center of gravity} , y _center of gravity are coordinates of the _center of gravity)
The difference from (Expression 1) is that the level value of the selected unprocessed pixel is not used as it is, but a level value that makes the total of the cells equal to the threshold value is used.

  This arithmetic expression is stored in the ROM 26 in advance as in the case of (Expression 1), and is executed by the CPU 22 reading it from the ROM 26 when executing this step. Here, the level value of the pixel finally selected in step S405 is calculated to be equal to the threshold value by taking the sum of the level values calculated so far, and the center of gravity position is calculated based on this. Will be asked. Then, the CPU 22 writes the calculated gravity center position data again in the working memory 272 of the RAM 27 (see the working memory 272 in FIG. 15B).

  Further, the CPU 22 calculates a value (error data) obtained by subtracting the calculation result calculated in step S409 from the original level value of the finally selected pixel, and writes the value in the input buffer of the selected pixel again. (Step S410). Thereby, the level value of the pixel is stored as return data. In the example shown in FIG. 15C, the value “10” is obtained by subtracting “50” calculated in step S409 from the level value “60” of the finally selected pixel (3, 2). The data is again written to the address (3, 2) of the input buffer 271 of the RAM 27 (see the input buffer 271 in FIG. 15C).

  Then, this pixel is not incorporated into the cell, but is processed as an unprocessed pixel incorporated into the cell in the initial scan (step S411). As described above, instead of returning the return data to the pixels other than the cells, the dot distribution faithful to the input gradation value can be formed by storing the return data again in the pixel where the return data is generated.

  Thereafter, the process proceeds to step S408, and the above-described process is repeated. Incidentally, the CPU 22 stores “255” in the output buffer 273 corresponding to the pixel where the center of gravity exists. Thereby, CPU22 can hit a dot to the pixel in which a gravity center position exists by subsequent processing.

  As described above, according to the present embodiment, after the gradation data of each color component is given to the input image by the pre-gradation conversion process by the error diffusion method, the post-gradation conversion process by the CC method is performed. Then, binarization processing is performed using a predetermined level value as a threshold value. By doing so, it is possible to obtain a visually comfortable output image in which generation of a periodic or regular pattern that does not exist in the input image or a peculiar pattern due to the bias of dots is suppressed. Further, according to the post-gradation conversion processing by the CC method, amplification of quantization error data that can occur in the pre-gradation conversion processing can be suppressed, so that an output image faithful to an input image in which generation of a unique pattern is suppressed can be obtained. You can get it.

  FIG. 16 shows, as an example, a case where the present embodiment is applied to an input image having a uniform density whose level value of gradation data of the input image is “46”. FIG. 16A shows the input image, FIG. 16B shows the output image when the post-gradation conversion process by the CC method is performed without performing the pre-gradation conversion process, and FIG. 16C shows the pre-gradation. The input image after the conversion processing is shown, and (d) shows the output image when the post-tone conversion processing by the CC method is performed after the pre-tone conversion processing. In FIG. 16B, the occurrence of a regular repetitive pattern is seen in the upper left part of the screen, but such a unique pattern does not exist in FIG. 16D. As apparent from the comparison between FIG. 16B and FIG. 16D, the combination processing of the pre-gradation conversion processing and the post-gradation conversion processing according to the present embodiment suppresses the occurrence of a singular pattern. A comfortable output image is obtained.

  As mentioned above, although one embodiment of the present invention has been described, the present invention is not limited to such an embodiment, and can of course be implemented in various forms without departing from the spirit of the present invention. is there. Hereinafter, a modification will be described.

"First modification"
Next, a first modification of the present invention will be described. In the above-described embodiment, the pixel selection reference point is set to the center of gravity position and updated at the time of pixel selection. However, in the first modification, the pixel selection reference point is not related to the level value of the pixel. The center position using the coordinate position of. This makes it possible to increase the overall processing speed including other processes without calculating the center of gravity position every time.

  A first modification will be described with reference to FIG. Here, the center position is not the position of the center of gravity using the level value but the center position using the coordinate position, and the position coordinate is (x, y) and is calculated by the following formula (the following Where N is the number of pixels in the cell).

(Formula 3)
x = Σ _i x _i / N
y = Σ _i y _i / N
A case will be described in which each pixel and the gradation value shown in FIG. 17A are input as input image data. FIG. 17A is an example in which image data for three vertical pixels and four horizontal pixels is input for the sake of simplicity, and even when image data for one frame is input. Good.

  When the input image data shown in FIG. 17A is input, an initial pixel is first selected, but the uppermost leftmost pixel among unprocessed pixels is selected as a predetermined scanning method as in the above embodiment. . This is also because the probability that the finally generated cells are arranged obliquely in one image is increased, and the dots are hit in an oblique direction and are difficult to perceive. In the example of FIG. 17A, the pixel at the position (n, m) is selected as the initial pixel.

  Then, when the initial pixel is incorporated into the cell CL and the center position is calculated using (Equation 3), the center position SC is located at the position shown in FIG. Next, an unprocessed pixel located closest to the center position SC is selected. When there are a plurality of such pixels, they are selected by a predetermined scanning method as in the above-described embodiment. In the example shown in FIG. 17B, there are two positions (n + 1, m) and (n, m + 1) closest to the center position SC. Here, (n + 1, m) is selected.

  Next, as shown in FIG. 17C, when the selected unprocessed pixel MG is incorporated into the cell CL and the cell enlargement process is performed, a reference point, that is, a center position for selecting the unprocessed pixel is calculated. As shown in FIG. 17 (c), the position is SC. Since the total level value of the cell is “100” at this time, an unprocessed pixel is further selected. Also in this modification, the threshold value is set to “255” as in the above embodiment, and unprocessed pixels are selected until this value is reached.

  The pixel closest to the reference point (center position) SC is selected. In the example shown in FIG. 17C, however, there are two pixels (n, m + 1) and (n + 1, m + 1). . In this case, the selection is made according to the scanning method, and the (n, m + 1) pixel is selected as an unprocessed pixel.

  Then, as shown in FIG. 17D, when the selected unprocessed pixel is incorporated into the cell and the center position is calculated using (Equation 3), it is located at the position SC. The total level value of the cell at this point is “165”, and the threshold value is not reached, so that an unprocessed pixel is selected. Since the pixel closest to the center position SC is the pixel located at (n + 1, m + 1), this pixel is selected as an unprocessed pixel.

  Next, when the selected unprocessed pixel MG is incorporated into the cell and the center position SC is calculated using (Equation 3), the position shown in FIG. The sum of the cell level values at this point is “215”, which does not reach the threshold value. Therefore, an unprocessed pixel is selected. The unprocessed pixel closest to the center position SC has four pixels (n, m + 2), (n + 1, m + 2), (n + 2, m), and (n + 2, m + 1), but according to a predetermined scanning method. The pixel located at (n + 2, m) is selected.

  The selected unprocessed pixel is incorporated into the cell, but the total level value of the cells generated when the unprocessed pixel is incorporated is “265”, which exceeds the threshold value. In this case as well, as in the previous embodiment, the level value corresponding to the threshold value is incorporated into the cell, the remaining level value (error data) is returned to the unprocessed pixel, and the pixel is then selected as the unprocessed pixel. So that This is to generate dots that are faithful to the gradation data of the input image, as in the above embodiment. Here, “10” is given as return data to the pixel located at (n + 2, m) (see FIG. 17F).

  Next, since the total level value of the cells has reached the threshold value, the cell configured so far performs a process for generating dots at the center of gravity of the cell. The center-of-gravity position is calculated in the same manner as (Formula 1) of the above-described embodiment, and the center-of-gravity position GC in FIG. 18A is obtained.

  Then, when a dot is generated in the pixel located at the center of gravity position GC by the subsequent processing, the dot is generated at the position shown in FIG. As in the previous embodiment, if the center of gravity position is calculated each time, the processing calculation increases as shown in (Expression 1) and (Expression 2), but the reference point for pixel selection is calculated as the center position, and finally the center of gravity is calculated. , The amount of calculation can be reduced, and the quantized pixel group can grow in a circle to maintain a dot-to-dot distance and obtain a visually comfortable dot output.

"Second modification"
Further, as a second modification, without calculating the center of gravity position of the cell, in the first modification, a dot is generated at the pixel at the center position of the cell used for selecting the unprocessed pixel. Also good. In this way, it is not necessary to calculate the center of gravity, so that the amount of calculation can be reduced. In this case, the central position may be located between a plurality of pixels. In such a case, for example, any one pixel in the cell may be selected according to the same priority order as the scanning method. For example, in the example shown in FIG. 18A, as shown in FIG. 18C, dots are generated at pixel positions different from those in FIG. Or you may make it select at random or select any one pixel from several pixels using a predetermined | prescribed table.

“Third Modification”
Furthermore, as a third modification, instead of updating the reference point for pixel selection every time, it is updated every other time, updated every second time, or the number of times is determined in advance and updated at that time. You may make it do. An example of updating every other time will be described with reference to FIG.

  As in the case of FIG. 16, the case where the input image data is input as shown in FIG. 19A will be described. Similarly to FIG. 17B, when the initial pixel is selected and its center position SC is calculated, the position shown in FIG. 19B is obtained. When a pixel closest to the unprocessed pixel MG from the reference point (center position SC) is selected, the pixel is located at (n + 1, m).

  Next, the selected unprocessed pixel is incorporated into the cell. The unprocessed pixel to be selected next is not selected based on the center position SC calculated from the cell, but is selected based on the center position SC before incorporation. Do not update the reference point at this point. As shown in FIG. 19C, the pixel closest to the center position SC is (n, m + 1), and this pixel is selected as the unprocessed pixel MG.

  Next, a cell incorporating the selected unprocessed pixel MG is as shown in FIG. Since the reference point is updated every other time, if the center position SC of the cell configured at this time is calculated, the center position SC moves to the position shown in FIG.

  Next, with reference to the updated center position, the unprocessed pixel MG closest to this position is selected and incorporated into the cell CL as shown in FIG. 19 (e). Then, without updating the reference point, the pixel closest to the center position SC calculated immediately before is selected as the unprocessed pixel MG (the pixel at the position (n + 2, m) is selected) to form a cell.

  As in the example of FIG. 17, since the total level value of the pixels in the cell incorporating the selected pixel exceeds the threshold, the excess is returned to the selected pixel as return data. When the center of gravity position for dot generation is calculated, a position GC shown in FIG. 19F is obtained, and a dot is generated at this position. When the reference point is updated, the reference point SC moves to the position shown in FIG.

  In this way, instead of updating the reference point for pixel selection every time, the calculation amount can be further reduced by updating every other time, and the processing speed of the entire image processing apparatus including other processing can be improved. Can be planned. Furthermore, the reference point may be updated when the unprocessed pixel is selected for the fourth time from the selection of the initial pixel, and then updated at the number of times so as to be subtracted from the third time, the second time,.

"Fourth modification"
In the above-described embodiment, the error diffusion method is used for pre-gradation conversion to vary the level value of the gradation data. However, as a fourth modification, a systematic dither method or an average error minimum method may be used. These methods are generally used as a binarization method for reproducing the gradation of a grayscale image. In the above-described embodiment, the color correction is performed after the pre-gradation conversion by the error diffusion method. However, the present invention is not limited to this, and the color correction or the color conversion processing is not performed. A pre-gradation conversion process may be performed as a process for providing the above.

  The systematic dither method uses a conversion table called a threshold matrix for the level value of the gradation data of each pixel, and compares the threshold value stored in the table. The level value of the input pixel is converted into gradation data to a predetermined level value depending on whether the level value of the input pixel is larger or smaller than the threshold value in the table corresponding to the pixel. By doing so, it is possible to give variation to the level value of the input pixel as in the above-described embodiment.

  The minimum average error method is a method of correcting the gradation data value of the next unprocessed pixel with a weighted average value of error data generated in the peripheral gradation data converted pixels. On the other hand, the error diffusion method used in the post-gradation conversion in the present embodiment diffuses error data generated during gradation data conversion of a certain pixel to surrounding unprocessed pixels that have not yet been converted to gradation data. It is something to add. Accordingly, the minimum average error method and the error diffusion method can be considered to be completely equivalent except for the handling at the image end, and the level value of the input pixel can be varied as in the above embodiment.

  Further, in the pre-gradation conversion in the embodiment and the fourth modification example, as is clear from the above description, the error data increases as the number of level values decreases, so that the variation width can be increased. Therefore, as a modification of the fourth modification, the type of level value may be varied in accordance with the level value of the gradation data of the input image at the time of pre-gradation conversion. For example, in the case of an image with a relatively low density where the level value of the input pixel is 127 or less, the number of types of level values after gradation conversion is increased, and an image with a relatively high density where the level value of the input pixel is 128 or more. In this case, the number of types of level values after gradation conversion is reduced. Or, conversely, the type of level value may be reduced in the case of an image with a low density, and the type of level value may be increased in the case of an image with a high density. In this way, the size of the error data can be controlled in accordance with the density of the input image, and the probability that the shape of the cell to be generated will change in accordance with the density area of the input image during post-gradation conversion increases. It can contribute to suppression.

“Fifth Modification”
In the above embodiment, the binarized data obtained by gradation conversion into two kinds of level values “0” and “255” by post gradation conversion is stored in the output buffer 273. This is because it is assumed that dots having a size corresponding to the level value “255” are hit in the subsequent dot forming process. In the fifth modified example, dots of a plurality of types of sizes may be hit, and post-gradation conversion may be performed to a plurality of types of level values corresponding to the plurality of types of sizes of dots.

  For example, the pulse width modulation unit 35 generates drive data by changing the pulse width of the drive pulse for driving the laser corresponding to the level value binarized by the post-gradation conversion process. Then, this drive data is output to the print engine 40, and the laser diode 42 is driven via the laser driver 41. As a result, since the dot size can be controlled by the driving time (on time) of the laser diode 42, dots having a size corresponding to the binarized level value are printed on a recording medium such as printing paper. Will be. In this case, there are as many types of level values output as binarized data after post-gradation conversion as the number of threshold types used for binarization. Therefore, the number of kinds of binarized level values is usually small compared to the kind of level values of gradation data before post-gradation conversion. However, the drive data is set by finely setting the pulse width described above. There are times when it will increase when generating.

  Actually, in the post-gradation conversion process, the gradation data of the input image is divided into the gradation data of the input image corresponding to a plurality of dots of a target size, and the plurality of dots are converted. It is also possible to perform binarization processing for forming. Since this processing is not the essence of the present invention, the description is omitted here.

“Sixth Modification”
As a sixth modification, the image processing apparatus in the above embodiment is configured by a laser printer. However, the present invention is not limited to this, and ink such as an ink jet printer is attached to a print medium. The printer may be configured by various printers such as a printer or a thermal printer, an electronic photograph or a display. Alternatively, the image processing apparatus of the present invention may be configured in various apparatuses such as a copier, a computer, a word processor, and a fax machine in which such a function as a printer is incorporated.

The block diagram which shows the whole system in one Embodiment with which this invention is applied. 1 is a schematic diagram illustrating a hardware configuration of an image processing apparatus according to an embodiment. Explanatory drawing explaining the mode of gradation conversion in gradation data conversion processing. 6 is a flowchart for explaining processing performed by the image processing apparatus according to the present embodiment. Explanatory drawing which shows the scanning method in the gradation conversion process of input image data. Explanatory drawing explaining the weighting dispersion | distribution method of the error data in an error diffusion method. 6 is a flowchart for explaining error diffusion processing in pre-gradation conversion processing. Explanatory drawing explaining the dispersion | distribution process of the gradation data in an error diffusion method. The flowchart explaining the post gradation conversion process by CC method. Explanatory drawing for demonstrating the process of the gradation conversion process in CC method. Explanatory drawing for demonstrating the process of the gradation conversion process in CC method. Explanatory drawing explaining the difference of the post gradation conversion by the presence or absence of a pre gradation conversion process. Explanatory drawing explaining the structure of RAM and the example of the data stored. Explanatory drawing explaining the structure of RAM and the example of the data stored. Explanatory drawing explaining the structure of RAM and the example of the data stored. The schematic diagram which showed an example of the output image obtained by the gradation conversion process of this embodiment. Process explanatory drawing by CC method in case the reference point of pixel selection is made into the center position of a cell. Process explanatory drawing by CC method in case the reference point of pixel selection is made into the center position of a cell. Process explanatory drawing by CC method in case the reference point of pixel selection is made into the center position of a cell.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Host computer, 11 ... Application part, 12 ... Rasterization part, 20 ... Image output device, 21 ... Input I / F, 22 ... CPU, 24 ... Total, 25 ... Hard disk, 26 ... ROM, 27 ... RAM, 30 ... Image processing unit 31 ... Pre-gradation conversion unit 32 32 Color conversion unit 33 Color correction table 34 Post-gradation conversion unit 35 Pulse width modulation unit 40 Print engine 41 Laser driver 42 ... Laser diode, 271 ... Input buffer, 272 ... Working memory, 273 ... Output buffer, CL ... Cell, CL1, CL2, CL3 ... Cell, GC ... Center of gravity of cell, GC2, GC3 ... Center of gravity of cell, SC ... Cell Center position, MG ... unprocessed pixel.

Claims (15)

For each pixel, gradation data having at least P (P is an integer of 2 or more) type level values is converted into gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. The first gradation data conversion is performed to convert gradation data having N (N is an integer of 2 or more) types of gradation values into gradation data having M (M is an integer of 2 or more) types of level values. An image processing apparatus that performs gradation data conversion of 2;
In performing the second gradation data conversion,
A pixel group generation unit that generates a pixel group by selecting pixels having the N types of level values until the sum of level values for each pixel is equal to or greater than a threshold;
A data providing unit that provides the M kinds of level values to the predetermined pixels of the generated pixel group;
An image processing apparatus. The image processing apparatus according to claim 1,
In performing the second gradation data conversion,
Furthermore, a pixel group centroid determining unit that determines the centroid position of the pixel group generated by the pixel group generating unit,
The pixel group generation unit generates the pixel group by continuing to select the pixel closest to the center of gravity determined by the pixel group center of gravity determination unit until the sum of level values for each pixel is equal to or greater than a threshold, and the data The assigning unit assigns the M kinds of level values to pixels located at the center of gravity determined by the pixel group center of gravity determination unit. The image processing apparatus according to claim 2,
The pixel group generation unit continues to select the pixel closest to the centroid determined by the pixel group centroid determination unit until the sum of the level values for each pixel is equal to or greater than a threshold, and the sum of the level values in the selected pixel group When the threshold value exceeds the threshold, the level value exceeding the threshold value is returned to the last pixel selected by the pixel group generation unit. The image processing apparatus according to any one of claims 1 to 3,
An image processing apparatus using a systematic dither method, an error diffusion method, or an average error minimum method as a method of gradation data conversion when performing the first gradation data conversion. For each pixel, gradation data having at least P (P is an integer of 2 or more) type level values is converted into gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. The first gradation data conversion is performed to convert gradation data having N (N is an integer of 2 or more) types of gradation values into gradation data having M (M is an integer of 2 or more) types of level values. 2 is an image processing method for performing gradation data conversion of 2;
In performing the second gradation data conversion,
A pixel group generation step of generating a pixel group by selecting pixels having the N types of level values until the sum of level values for each pixel is equal to or greater than a threshold;
A pixel group centroid determination step for determining a centroid position of the pixel group generated in the pixel group generation step;
A data providing step of assigning the M kinds of level values to pixels located at the center of gravity determined in the pixel group center of gravity determination step;
The pixel group generation step generates the pixel group by continuing to select a pixel closest to the center of gravity determined in the pixel group center of gravity determination step until a sum of level values for each pixel is equal to or greater than a threshold value. An image processing method. For each pixel, gradation data having at least P (P is an integer of 2 or more) type level values is converted into gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. The first gradation data conversion is performed to convert gradation data having N (N is an integer of 2 or more) types of gradation values into gradation data having M (M is an integer of 2 or more) types of level values. 2 is an image processing program for performing gradation data conversion of 2;
In performing the second gradation data conversion,
A pixel group generation process for generating a pixel group by selecting pixels having the N types of level values until the sum of level values for each pixel is equal to or greater than a threshold;
A pixel group centroid determination process for determining a centroid position of the pixel group generated by the pixel group generation process;
Causing the computer to execute a data addition process for assigning the M kinds of level values to the pixels located at the center of gravity determined in the pixel group center of gravity determination process;
In the pixel group generation process, the pixel group is generated by continuously selecting a pixel closest to the center of gravity determined in the pixel group center of gravity determination process until a sum of level values for each pixel becomes equal to or greater than a threshold value. Program.   A computer-readable recording medium on which the image processing program according to claim 6 is recorded. For each pixel, gradation data having at least P (P is an integer of 2 or more) type level values is converted into gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. The first gradation data conversion is performed to convert gradation data having N (N is an integer of 2 or more) types of gradation values into gradation data having M (M is an integer of 2 or more) types of level values. An image processing apparatus for performing gradation data conversion of 2;
In performing the second gradation data conversion,
A pixel group generation unit that generates a pixel group by selecting pixels having the N types of level values until a sum of level values for each pixel is equal to or greater than a threshold value based on a first reference point for pixel selection; ,
A pixel group reference point determination unit for determining a second reference point of the pixel group generated by the pixel group generation unit;
A data providing unit that provides the M kinds of level values to the pixels located at the second reference point determined by the pixel group reference point determining unit;
The image processing apparatus, wherein the pixel group generation unit updates the first reference point and selects a pixel based on the first reference point. The image processing apparatus according to claim 8.
The image group generation unit updates the first reference point every time the pixel is selected a predetermined number of times. The image processing apparatus according to claim 8.
The first reference point is a center position calculated from the position of the selected pixel, and the second reference point is a center of gravity position calculated from the pixel position of the pixel group and the gradation pixel data value. An image processing apparatus comprising: The image processing apparatus according to claim 8.
The first reference point is a center position calculated from the position of the selected pixel, and the second reference point is a center position calculated from the pixel position of the pixel group. Processing equipment. The image processing apparatus according to any one of claims 8 to 11,
An image processing apparatus using a systematic dither method, an error diffusion method, or an average error minimum method as a method of gradation data conversion when performing the first gradation data conversion. For each pixel, gradation data having at least P (P is an integer of 2 or more) type level values is converted into gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. The first gradation data conversion is performed to convert gradation data having N (N is an integer of 2 or more) types of gradation values into gradation data having M (M is an integer of 2 or more) types of level values. 2 is an image processing method for performing gradation data conversion of 2;
In performing the second gradation data conversion,
A pixel group generation step of generating a pixel group by selecting pixels having the N types of level values until the sum of the level values for each pixel becomes equal to or greater than a threshold value based on a first reference point for pixel selection; ,
A pixel group reference point determining step for determining a second reference point of the pixel group generated in the pixel group generating step;
A data providing step of assigning the M kinds of level values to the pixels located at the second reference point determined in the pixel group reference point determining step,
In the pixel group generation step, the first reference point is updated and a pixel is selected based on the first reference point. For each pixel, gradation data having at least P (P is an integer of 2 or more) kinds of level values is converted to gradation data having Q (Q <P, P is an integer of 2 or more) kinds of level values. First gradation data conversion is performed to convert gradation data having N (N is an integer of 2 or more) types of gradation values into gradation data having M (M is an integer of 2 or more) types of level values. 2 is an image processing program that performs gradation data conversion of 2;
In performing the second gradation data conversion,
A pixel group generation process for generating a pixel group by selecting pixels having the N kinds of level values until the sum of the level values for each pixel is equal to or greater than a threshold value based on the first reference point for pixel selection; ,
A pixel group reference point determination process for determining a second reference point of the pixel group generated by the pixel group generation process;
Causing the computer to execute a data adding process for adding the M kinds of level values to the pixels located at the second reference point determined in the pixel group reference point determining process;
The pixel group generation process updates the first reference point and selects a pixel based on the first reference point. A computer-readable recording medium on which the image processing program according to claim 14 is recorded.