JP5669249B2 - Program, image processing apparatus, and image processing method - Google Patents

Description

  The present invention relates to a program, an image processing apparatus, and an image processing method.

  In order to output an image with a printer, it is usually necessary to perform quantization processing on the image. The quantization process is a halftone process for converting an image expressed in continuous gradation into the number of gradations that can be expressed by a printer. Conventionally, for example, dither processing and error diffusion processing are known as quantization processing methods.

  Conventionally, a method of performing quantization by selectively using dither processing and error diffusion processing is known (see, for example, Patent Documents 1 and 2). For example, in the method disclosed in Patent Document 1, dither processing and error diffusion processing are selectively used depending on the density range. In the method disclosed in Patent Document 2, a print head capable of forming N types of dots having different sizes is used, and dither processing and error diffusion processing are selectively used according to the dot size.

Japanese Patent Laid-Open No. 2006-240054 JP 2008-87382 A

  The dither processing is a method of applying a dither matrix value (noise) to pixel data and determining the presence or absence of halftone dots, and performing quantization by paneling the dither matrix to the original image. For this reason, there may be a texture problem in which a specific pattern generated in the matrix is repeatedly generated. As a result, image quality may be degraded.

  For example, in order to prevent the occurrence of a texture peculiar to dither processing, it is conceivable to adjust the value (noise) of the dither matrix to an appropriate value. However, suppressing the generation of texture by such adjustment requires a lot of work time and is difficult in practical use.

  On the other hand, in the error diffusion processing, a threshold value to be compared with the pixel data is prepared, and the pixel data value corrected by the accumulated error calculated in advance is compared with the threshold value, and the presence / absence of halftone dots is determined depending on the magnitude. It is a method of determination. When the error diffusion process is used, an error caused by quantization can be reduced in the entire image, so that a texture that occurs in the dither process does not occur.

  However, when error diffusion processing is used, dot delays that cause dot placement to be delayed in the output results at the highlight and shadow portions, worm noise that causes dots to be linked, and dot placement that is part of the halftone portion In particular, there may be a problem of pattern noise that is randomly arranged with a checkered pattern. As a result, image quality may be deteriorated due to an image shift due to dot delay, a stripe due to worm noise, or the like.

  For example, when dither processing and error diffusion processing are separately used as in Patent Document 1, a boundary line due to a difference in quantization processing method appears at a portion where the dither processing is switched to error diffusion processing, and the image quality is low. There is a risk of deteriorating problems. Further, when the method as in Patent Document 2 is used, for example, since error diffusion processing is used in a highlight portion, there is a possibility that a problem or the like that leads to degradation of image quality due to occurrence of dot delay may occur.

  Therefore, conventionally, it is desired to perform quantization processing by a more appropriate method. More specifically, for example, a quantization process capable of realizing high-quality halftone processing by suppressing generation of texture, dot delay, worm noise, pattern noise, and the like is desired.

  In recent years, an inkjet printer or the like that performs multi-tone printing by ejecting a plurality of types of ink droplets having different dot sizes may be used in accordance with an increase in required print quality. In this case, in the quantization process, for example, it is necessary to perform multi-tone quantization corresponding to a plurality of types of dot sizes. In this case, for example, it is desired to perform the quantization process by a more appropriate method applicable to multi-gradation quantization. SUMMARY An advantage of some aspects of the invention is that it provides a program, an image processing apparatus, and an image processing method that can solve the above-described problems.

  The inventor of the present application, based on diligent research, has focused on the fact that the ease of texture generation differs depending on the pixel density in dither processing. In the error diffusion method, attention was also paid to the fact that dot delay and worm noise are likely to occur depending on the pixel density. Then, for example, instead of switching between dither processing or error diffusion processing and performing one processing, quantization is performed by allocating the influence of the dither processing and error diffusion processing, so that texture, dot delay, and worm noise are performed. It was found that quantization can be appropriately performed while suppressing the occurrence of the above.

  In addition, the inventors of the present application have further conducted intensive studies and found that this method is also effective when performing multi-tone quantization. The present invention has the following configuration.

(Configuration 1) A program for causing a computer to perform multi-gradation quantization processing, which is processing for converting density values indicating the color density of each pixel in an image into discrete values of three or more stages, Pixel selection processing for sequentially selecting pixels to be quantized, noise usage rate indicating the degree of influence of dither matrix noise, which is noise specified by a preset dither matrix, on quantization processing, and surrounding pixels Is a process for determining an error usage rate indicating the degree to which the accumulated error obtained by accumulating the quantization error, which is an error occurring in the quantization with respect to the quantization process, is affected, and the density value of the pixels sequentially selected in the pixel selection process The usage rate determination process for determining the noise usage rate and the error usage rate corresponding to the pixel and the density value of the pixels sequentially selected in the pixel selection process are quantized. A quantization execution process that uses dither matrix noise according to the noise usage rate corresponding to the pixel and performs quantization using the accumulated error according to the error usage rate corresponding to the pixel; the cause the computer to execute, usage determination process, even when the concentration value of the pixel is either the noise utilization, set to a preset minimum noise utilization value greater than or equal to a value greater than 0, the highlight The noise usage rate is gradually reduced and the error usage rate is gradually increased with respect to the change in the pixel density value in the direction from the highlight portion to the halftone portion between the halftone portion and the halftone portion. The noise usage rate is gradually increased and the error usage rate is gradually reduced with respect to the change in the density value of the pixel in the direction from the halftone portion to the shadow portion between the halftone portion and the shadow portion. To do . Multi-tone quantization is, for example, multi-value quantization processing for performing multi-value quantization on a pixel density value.

  The dither processing and the error diffusion processing are different in density range in which a high quality image can be obtained. Therefore, when any one of these is used alone, there is a possibility that the image quality is deteriorated due to quantization in a part of the density range. Further, for example, if the dithering process and the error diffusion process are simply switched depending on the density range, there is a possibility that a boundary line or the like is generated at a part where the process is switched.

  On the other hand, if configured in this way, for example, the spatial frequency characteristic of the error diffusion process (hereinafter referred to as error diffusion characteristic) can be affected by the spatial frequency characteristic of the dither process (hereinafter referred to as dither characteristic). In addition, this enables, for example, quantization to be performed by a method in which a dither characteristic is incorporated into the error diffusion characteristic. The spatial frequency characteristic is, for example, a characteristic when one of the characteristics for evaluating the image quality is a repetition pattern of ON / OFF of an output result as a frequency.

  Furthermore, in the case of such a configuration, unlike the case of simply switching between dither processing and error diffusion processing, for example, it is possible to perform quantization processing that makes better use of the portion that each processing is good at. Therefore, with this configuration, for example, various problems relating to printing that occur in the quantization process can be appropriately solved, and the quantization process can be performed by a more appropriate method.

  Also, with this configuration, even when multi-tone quantization processing is performed, each process of dither processing and error diffusion processing is performed without causing a noticeable boundary line or the like at a portion where processing is switched. Thus, it is possible to perform a quantization process that makes better use of the portion that is good at. Therefore, with this configuration, for example, it is possible to perform multi-tone quantization processing by a more appropriate method.

  In the pixel selection process, for example, pixel lines are sequentially selected, and pixels in the selected line are sequentially selected along a predetermined processing direction. In this case, in the pixel selection process, the processing direction may be switched for each line to be processed. For example, it is conceivable to sequentially select pixels from left to right in odd lines and from right to left in even lines. With this configuration, for example, the quantization process becomes a bidirectional process, and the error diffusion direction is not constant, so that the dots can be more appropriately dispersed.

  In the usage rate determination process, for example, a noise usage rate and an error usage rate are determined according to a preset calculation formula. In this case, for example, by appropriately setting the parameters in the equation, it is possible to appropriately suppress the occurrence of a boundary line at the switching portion between the area where the dither characteristic is dominant and the area where the error diffusion characteristic is dominant. Can do. In addition, this enables smooth switching of the quantization processing method. The optimum value of the parameter can be appropriately obtained by experiments or the like, for example.

  Also, this program causes a computer to execute each process for each process color, for example. In this case, for example, the dither matrix to be used may be changed for each process color. With this configuration, for example, the value of the dither matrix used in quantization for the same pixel can be made different for each process color. Moreover, it can prevent appropriately that a dot overlap arises by this.

  As the dither matrix noise, for example, it is preferable to use blue noise characteristics. The blue noise characteristic is, for example, noise biased to a high frequency when expressed by a spatial frequency characteristic, and has a characteristic that is difficult to perceive by human vision. The blue noise characteristic is similar to the frequency characteristic of error diffusion processing. Therefore, by using such a dither matrix, it becomes easier to smoothly switch between the dither characteristic and the error diffusion characteristic.

  Further, this program may cause the computer to further perform, for example, processing for calculating an accumulated error. For example, this program may cause the computer to further execute error distribution processing and the like. The error distribution process is a process for distributing a quantization error caused by pixel quantization, for example, and diffuses the quantization error to pixels around the pixel according to a preset diffusion filter (diffusion matrix), for example. Let In this way, the value of the accumulated error corresponding to each of the surrounding pixels is updated. As the diffusion filter, for example, a matrix of Jarvis, Judice & Ninke can be suitably used.

  (Configuration 2) As for discrete values of three or more levels, the minimum quantized value corresponding to the result of quantizing the minimum density value and a value larger than the minimum density value by a predetermined amount or more are quantized. And at least two quantization result values that are discrete values larger than the minimum quantization value, and the quantization execution process includes two or more quantization result values. It is processing corresponding to each, and includes step-by-step determination processing for determining whether or not the result of quantizing the density value should be a value greater than or equal to the quantization result value. The step-by-step determination process performed corresponding to at least one uses dither matrix noise according to the noise usage rate corresponding to the pixel to be processed, and according to the error usage rate corresponding to the pixel. Use cumulative error to find minimum quantum Quantization processing of two gradations between two gradations of the value and the quantization result value is performed, and quantization execution processing is determined for each stage corresponding to each of two or more quantization result values Based on the result of the processing, one of the discrete values of three or more levels is determined as the result of multi-gradation quantization for the density values of the pixels that are sequentially selected in the pixel selection processing. The lowest quantized value is, for example, the value 0. In this case, the quantization result values of two or more steps are, for example, discrete values of two or more steps larger than 0.

  With this configuration, for example, in multi-tone quantization processing, it is possible to perform processing that appropriately utilizes the respective characteristics of dither processing and error diffusion processing. This also makes it possible to appropriately perform multi-tone quantization processing.

  Note that determining whether or not the result of quantizing the density value should be a value greater than or equal to the quantization result value is, for example, determining whether or not the quantization result should be equal to the quantization result value. Alternatively, it may be determined whether or not the quantization result should include a value other than the quantization result value itself and be a value greater than or equal to the quantization result. In the latter case, the step-by-step determination processing is performed, for example, when the quantization result value is not the corresponding quantization result value itself but the larger quantization result value should be the quantization result. You may determine with the above value.

  For example, the step-by-step determination process corresponding to each quantization result value may be changed according to the result of the step-by-step determination process corresponding to the quantization result value at another step. For example, if the step-by-step determination process corresponding to a larger quantization result value (hereinafter referred to as an upper result value) determines that the quantization result should be greater than or equal to the upper result value, a smaller quantization In the determination processing for each stage corresponding to the result value (hereinafter referred to as the lower result value), for example, a quantization process using a dither process and an error diffusion process may be omitted.

  In this case, since it is known that the quantization result is greater than or equal to the higher result value than the lower result value, the step-by-step determination process corresponding to the lower result value is, for example, the second order for the lower result value. While omitting the key quantization process, it may be determined that the quantization result should be equal to or higher than the lower result value.

  More specifically, the quantization process is preferably performed as follows, for example. In the quantization execution processing, for example, the following processing is performed as processing in the case of performing two-tone quantization processing in each stage determination processing.

  For example, the program is a process of calculating an error-corrected input value, which is a density value after correction by cumulative error, for pixels sequentially selected in the pixel selection process, and an error usage rate corresponding to the pixel An error corrected input value calculation process for calculating a value obtained by adding the product of the accumulated error and the density value of the pixel as an error corrected input value, and a threshold value reflecting dither matrix noise as a threshold value used in quantization Is a process for calculating a noise-corrected threshold value, and a noise correction value for calculating a value obtained by adding the product of the noise usage rate corresponding to the pixel and the dither matrix noise to a preset initial threshold value as the noise-corrected threshold value The computer further executes the completed threshold value calculation process, and the quantization execution process performs the quantization by comparing the noise corrected threshold value and the error corrected input value.

  In such a configuration, the degree of influence of the error diffusion characteristic can be appropriately adjusted by adding the accumulated error multiplied by the error usage rate to the density value of the pixel which is the quantization input value. Further, by adding the dither matrix noise multiplied by the noise usage rate to the initial threshold value, the degree of influence on the dither characteristics can be adjusted appropriately.

  Therefore, with this configuration, for example, the degree of influence of each of the error diffusion characteristic and the dither characteristic can be appropriately set according to the input value. In addition, according to the input value, it is possible to perform a quantization process that more appropriately utilizes a portion that each process is good at.

  Further, for example, the quantization execution process includes a maximum value determination process for determining whether or not a pixel density value, which is an input value for quantization, is equal to a maximum value in a range that the density value can take, and an input value. A minimum value determination process for determining whether a density value of a pixel is equal to a minimum value in a range that the density value can take, and a quantization value acquisition process for acquiring a quantization value as a result of quantization If the maximum value determination process determines that the input value is equal to the maximum value, the quantization value acquisition process acquires a value to be output when the density value is greater than the threshold value as the quantization value. When it is determined that the input value and the minimum value are equal in the minimum value determination process, the quantized value acquisition process acquires a value to be output when the density value is smaller than the threshold value as the quantized value. .

  For example, if the input value is equal to the maximum value, regardless of whether the error-corrected input value or the noise-corrected threshold value, the quantization result is an output that should be output when the density value is greater than the threshold value. It needs to be a value (for example, 1). In addition, when the input value is equal to the minimum value, the quantization result is an output to be output when the density value is smaller than the threshold value, regardless of the error-corrected input value or the noise-corrected threshold value. It needs to be a value (for example, 0).

  On the other hand, if comprised in this way, the output value in case an input value is the maximum value or the minimum value can be set easily and appropriately. This also makes it possible to more appropriately perform the quantization process using both the accumulated error and the dither matrix noise. The output value to be output when the density value is larger than or smaller than the threshold value indicates, for example, a result of comparing the density value with the threshold value when both the accumulated error and the dither matrix noise are set to 0. Output value.

  Further, for example, the usage rate determination process determines the noise usage rate when the density value of the pixel is a density value corresponding to either the highlight portion or the shadow portion in the determination of the noise usage rate and the error usage rate. When the density value is equivalent to a halftone part between the light part and the shadow part, the value is larger than the noise usage rate, and the density value of the pixel is a density value corresponding to either the highlight part or the shadow part. The error usage rate in a certain case is set to a value smaller than the error usage rate in the case of the density value corresponding to the halftone portion.

  For example, the noise usage rate has a value of 1 (100%) in the highlight portion and the shadow portion, and changes so that the value gradually decreases as the tone becomes halftone. For example, the error usage rate changes so that the value is 0 (0%) at both ends of the density range, and the value is 1 (100%) before the density range of the texture generation unit.

  In such a configuration, for example, in a highlight portion and a shadow portion where dot delay is likely to occur when the influence of the error diffusion characteristic is strong, the noise usage rate is increased and the error usage rate is decreased. Thereby, for example, in the highlight portion and the shadow portion, the influence of the dither characteristics that can disperse and arrange the dots becomes large, so that the occurrence of dot delay can be appropriately suppressed.

  Also, by increasing the influence of the error diffusion characteristic and performing the quantization process in the halftone part, for example, a more natural pseudo gradation can be obtained. Further, for example, in the density range of the texture generation unit, by changing the noise usage rate and increasing the error usage rate, the dot arrangement is changed, and for example, the occurrence of texture can be appropriately suppressed.

  For example, in the usage rate determination process, the noise usage rate is set to a value equal to or higher than the minimum noise usage rate preset to a value larger than 0 regardless of the density value of the pixel.

  In the halftone part, if the noise usage rate is 0 (0%) and the error usage rate is 1 (100%), there is a possibility that pattern noise peculiar to the error diffusion characteristic occurs. The cause of the pattern noise occurring in the halftone part is, for example, that the error diffusion characteristic in the halftone part is very high frequency, so that the dot arrangement tends to be a checkered pattern, and the dot arrangement pattern as in dither processing. This is because the checkerboard pattern may partially occur due to the fact that is not constant. On the other hand, by setting the minimum noise usage rate so that the noise usage rate does not become 0, for example, in the halftone part, the influence of the error diffusion characteristic is increased while slightly affecting the dither characteristic. Thus, since it is slightly affected by the dot arrangement pattern of the dither processing, the overall arrangement pattern is likely to be constant, and a partial checkered pattern is less likely to occur. Thereby, for example, pattern noise can be appropriately suppressed.

  In addition, for example, the usage rate determination process includes a first highlight reference value set in advance as a reference indicating the density range in the highlight portion, and a second highlight reference having a value larger than the first highlight reference value. And a first shadow reference value that is set in advance and a second shadow reference value that is larger than the first shadow reference value as a reference that indicates the density range in the shadow portion. As a reference indicating the density range in the center, a first halftone reference value that is larger than the second highlight reference value and smaller than the first shadow reference value, and larger than the first halftone reference value, and When the second halftone reference value smaller than the first shadow reference value is used and the pixel density value is less than or equal to the first highlight reference value or greater than or equal to the second shadow reference value, the noise usage rate is set to 1 Set to If the value is greater than or equal to the first halftone reference value and less than or equal to the second halftone reference value, the noise usage rate is set to the lowest noise usage rate, and the pixel density value is greater than or equal to the first highlight reference value, and In the case of the first halftone reference value or less, the noise usage rate is a value that is equal to or higher than the minimum noise usage rate and 1 or lower, and is gradually decreased from 1 according to the difference between the pixel density value and the first highlight reference value. When the pixel density value is not less than the second halftone reference value and not more than the second shadow reference value, the noise usage rate is not less than the minimum noise usage rate and not more than 1 and the pixel Is set to a value that is gradually increased from the lowest noise usage rate in accordance with the difference between the density value and the second halftone reference value, and the pixel density value is not less than the second highlight reference value and not more than the first shadow reference value. In this case, the error usage rate is set to 1, and the pixel density value is set to the second highlight basis. If the value is less than or equal to the value, the error usage rate is a value between 0 and 1, and is set to a value that gradually decreases from 1 depending on the difference between the second highlight reference value and the density value. When the reference value is equal to or greater than the first shadow reference value, the error usage rate is a value between 0 and 1 and is set to a value that is gradually decreased from 1 according to the difference between the pixel density value and the first shadow reference value.

  If comprised in this way, according to the density value of a pixel, each of an error usage rate and a noise usage rate can be set appropriately, for example. This also makes it possible to appropriately perform a quantization process that makes better use of the parts that the error diffusion process and the dither process are good at.

  Further, for example, the usage rate determination process further uses a third highlight reference value that is greater than 0 and less than or equal to the first highlight reference value as a reference indicating the density range in the highlight portion, and uses the third highlight reference value for the shadow portion. As a reference indicating a certain density range, a third shadow reference value that is equal to or larger than the second shadow reference value and smaller than the maximum value of the range that the density value can take is further used. If the value is less than the value, the error usage rate is set to 0, and if the pixel density value is not less than the third highlight reference value and not more than the second highlight reference value, the error usage rate is set to the density value and the third high value. The difference from the light reference value is set to a value obtained by dividing the difference between the second highlight reference value and the third highlight reference value, and the pixel density value is equal to or greater than the first shadow reference value and the third shadow reference value. In the following cases, the error usage rate is set to the third shadow. When the difference between the reference value and the density value is divided by the difference between the third shadow reference value and the first shadow reference value, and the pixel density value is greater than or equal to the third shadow reference value, the error usage rate is set to Set to 0. If comprised in this way, an error usage rate can be set more appropriately, for example.

  (Configuration 3) A program causes a computer to perform multi-tone quantization during image forming processing for forming printable data indicating an image in a format that can be interpreted by the printing apparatus, and the printing apparatus has different dot sizes. This is an inkjet printer that performs multi-tone printing by ejecting multiple types of ink droplets, and each of the two or more levels of quantization result values is associated with the respective dot size of the multiple types of ink droplets. The step-by-step determination process determines whether or not to form dots of ink droplets of the dot size corresponding to the step-by-step determination process by performing a two-gradation quantization process. If it is not determined that a dot is formed in the process, the quantization execution process determines the lowest quantization value as the result of multi-tone quantization, and the determination process for each stage If it is determined that a dot is to be formed, the quantization execution process uses the multi-level quantization of the quantization result value corresponding to the maximum dot size in the determination process for each stage determined to form a dot. Determine the result.

  If comprised in this way, multi-tone printing can be performed appropriately according to the result of multi-tone quantization, for example. Thereby, high-definition printing can be performed more appropriately.

  (Configuration 4) In the quantization execution process, the stage-by-stage determination process corresponding to a larger dot size is performed first, and in the stage-by-stage determination process corresponding to one of the dot sizes, the dot of the ink droplet of the dot size is When it is determined to form, the step-by-step determination process corresponding to a dot size smaller than the dot size determines not to form a dot regardless of the density value of the pixel. The step-by-step determination process corresponding to the dot size is, for example, a quantization execution process further corresponding to the quantization result value for the quantization result value corresponding to the dot size.

  In such a configuration, when it is determined that the stage determination process corresponding to any dot size forms a dot, the process performed in the subsequent stage determination process can be omitted or simplified. Therefore, with this configuration, multi-tone quantization can be performed more efficiently.

  In the case of such a configuration, for example, the processing is performed in order from the step-by-step determination processing corresponding to a larger quantization result value. Then, if it is not determined that any dot is formed in any of the determination processes performed at each stage, the determination process for each stage performs a two-tone quantization process, and the quantization result is converted into a corresponding quantum. It is determined whether or not it should be a conversion result value. When it is determined that the quantization result should be the quantization result value, it is determined that a dot is to be formed.

  In addition, when it is determined that a dot is to be formed in any one of the determination processes performed in each step, the determination process for each step omits the two-gradation quantization process. In this case, this step-by-step determination process determines that the quantization result should be a quantization result value larger than the corresponding quantization result value based on the result of the step-by-step determination process performed previously. Accordingly, it is determined that no dot is formed for the dot size corresponding to the determination processing for each stage. With this configuration, for example, it is possible to appropriately determine that dots are formed only in one of the determination processes for each stage, and it is possible to appropriately determine that dots are not formed in the determination processes for other stages. it can.

  (Configuration 5) The program is a process of calculating an input value for each stage, which is a value used as an input value in the stage-by-stage determination process corresponding to each dot size, and sets the density value of the pixels sequentially selected in the pixel selection process. Accordingly, based on preset input value calculation data, the computer further executes an input value calculation process for calculating the input value for each stage corresponding to each dot size. The number of dots of ink droplets to be formed per unit area in accordance with the density value is determined for each dot size, and the input value calculation process uses the density value of the pixels sequentially selected in the pixel selection process. On the other hand, the input value for each stage corresponding to each dot size is represented by the number of dots associated with the density value for the dot size in the input value calculation data. Accordingly, when performing two-level quantization, the step-by-step determination processing corresponding to each dot size is performed by performing two-step quantization processing on the input value for each step corresponding to the dot size. Do. As the input value calculation data for each stage, for example, a dot size determination table indicating the ratio of dot sizes to be output for each density can be used.

  With this configuration, the dot size to be formed can be set more appropriately according to the density value of the pixel. This also makes it possible to perform multi-tone quantization more appropriately and appropriately print a high quality image.

  (Configuration 6) The quantization execution process further includes a stage-by-stage cumulative error calculation process for calculating a cumulative error corresponding to the quantization result value corresponding to each of two or more stages of quantization result values. Each determination process uses the accumulated error corresponding to the same quantization result value as the step-by-step determination process as the cumulative error. Based on the result of the process by the step-by-step determination process, The accumulated error corresponding to the quantization result value is updated.

  If comprised in this way, an appropriate accumulation error can be used in the process performed by each stage accumulated error calculation process, for example. In addition, for example, it is possible to more appropriately perform a quantization process that makes use of a portion that each process of the dither process and the error diffusion process is good at.

  (Configuration 7) An image processing apparatus that performs multi-gradation quantization processing, which is processing for converting density values indicating the color density of each pixel in an image into discrete values of three or more stages, A pixel selection processing unit that sequentially selects pixels to be quantized, a noise usage rate indicating a degree of influence of the dither matrix noise, which is noise specified by a preset dither matrix, on quantization processing, and surrounding pixels Is a processing unit that determines an error usage rate that indicates a degree to which a cumulative error obtained by accumulating a quantization error, which is an error that occurs in quantization with respect to the quantization process, affects the quantization process. A usage rate determination processing unit that determines a noise usage rate and an error usage rate corresponding to the pixel in accordance with the density value, and a quantization for the density value of pixels sequentially selected by the pixel selection processing unit. Quantization execution that uses the dither matrix noise according to the noise usage corresponding to the pixel and performs quantization using the accumulated error according to the error usage corresponding to the pixel A processing unit. If comprised in this way, the effect similar to the structure 1 can be acquired, for example.

  The image processing apparatus may be a computer that operates according to a predetermined program, for example. In this case, for example, the CPU of the computer operates as each unit of the image processing apparatus according to the program.

  (Configuration 8) An image processing method for performing multi-gradation quantization processing, which is processing for converting density values indicating the color density of each pixel in an image into discrete values of three or more stages, A pixel selection processing stage for sequentially selecting pixels to be quantized, a noise usage rate indicating the degree of influence of the dither matrix noise, which is noise specified by a preset dither matrix, on the quantization processing, and surrounding pixels This is a processing stage for determining an error usage rate indicating the degree to which the accumulated error obtained by accumulating the quantization error, which is an error occurring in the quantization with respect to the quantization process, affects the quantization process. According to the density value, the usage rate determination processing stage for determining the noise usage rate and the error usage rate corresponding to the pixel, and the density value of the pixels sequentially selected in the pixel selection processing stage Quantization is performed using the dither matrix noise according to the noise usage rate corresponding to the pixel and the accumulated error according to the error usage rate corresponding to the pixel. Performing a quantization execution processing stage. In this way, for example, the same effect as that of Configuration 1 can be obtained.

  According to the present invention, for example, multi-tone quantization processing can be performed by a more appropriate method. Accordingly, for example, in halftone processing, texture, dot delay, pattern noise, and the like can be appropriately suppressed.

FIG. 10 is a diagram illustrating an example of a program that performs a two-tone quantization process. FIG. 1A shows an example of the configuration of a printing system 10 that uses this program. FIG. 1B is a diagram illustrating an outline of quantization processing performed by the image processing apparatus 12. It is a flowchart which shows an example of the operation | movement which performs quantization. It is a figure explaining in more detail about pixel selection processing S102. FIG. 3A shows an example of the order in which pixels are sequentially selected. FIG. 3B is a diagram illustrating an example of the effect of bidirectional processing. An example of a graph and a calculation formula for associating the error usage rate Re, the noise usage rate Rn, and the density value In (x, y) is shown. It is a flowchart which shows an example of the process which calculates error utilization factor Re. It is a flowchart which shows an example of the process which calculates noise usage rate Rn. It is a flowchart which shows an example of operation | movement of quantization execution process S110. It is a figure which shows an example of the spreading | diffusion filter used in this example. FIG. 8A shows an example of a diffusion filter used in each direction of bidirectional processing. FIG. 8B shows an example of how to distribute errors. It is a flowchart which shows an example of operation | movement of error distribution process S112. An example of the printing result according to this example is shown together with the printing result according to the reference example. FIG. 10A shows an example of a print result when only a conventionally known typical error diffusion process is performed. FIG. 10B shows an example of a print result when only a conventionally known dither process is performed. FIG. 10C shows an example of a print result when the dither process and the error diffusion process are simply switched depending on the density range. FIG. 10D shows an example of the print result according to this example. It is a flowchart which shows the 1st example of the process of the multi-tone quantization which concerns on one Embodiment of this invention. It is a figure explaining in more detail about input value calculation processing S608 and input value threshold value correction processing S610. FIG. 12A is a graph showing an example of a dot size determination table used in the input value calculation process S608. FIG. 12B is a diagram for explaining the input value threshold value correction processing S610. It is a flowchart which shows an example of the specific process in input value threshold value correction process S610. It is a flowchart which shows an example of the process in quantization execution process S612. It is a flowchart which shows an example of the process in Large quantization process S642. It is a flowchart which shows the modification of the process in quantization execution process S612. It is a flowchart which shows the 2nd example of the process of multi-gradation quantization. It is a flowchart which shows an example of the process which correct | amends the calculated noise usage rate. It is a figure which shows an example of the range which corrects a noise usage rate. FIG. 19A shows an example of a dot size determination table. FIG. 19B shows another example of the dot size determination table.

  Hereinafter, embodiments according to the present invention will be described with reference to the drawings. First, prior to the multi-gradation quantization processing according to the present invention, the two-gradation quantization processing related to the multi-gradation quantization processing will be described with reference to FIGS.

  FIG. 1 is a diagram illustrating an example of a program that performs two-tone quantization processing. FIG. 1A shows an example of the configuration of a printing system 10 that uses this program. In this example, the printing system 10 includes an image processing device 12 and a printing device 14.

  The image processing apparatus 12 is an apparatus that performs image forming processing such as RIP (Raster Image Processor) processing. The image processing device 12 develops the original image indicated by the print data, thereby forming printable data indicating the image in a format interpretable by the printing device 14.

  In this example, the image processing apparatus 12 quantizes at least a density value indicating the color density of each pixel in the original image in the image forming process. In this case, the image processing apparatus 12 performs quantization for each process color used in the printing apparatus 14. Accordingly, the image processing apparatus 12 forms a halftone image corresponding to each process color based on the print data.

  The image processing apparatus 12 is a host PC that controls the printing apparatus 14, for example, and operates as an image processing apparatus in accordance with a predetermined program. The image processing apparatus 12 may receive print data from another PC, for example. The print data may be created on the image processing apparatus 12 by the user.

  The printing device 14 is, for example, an ink jet printer, and executes image printing according to printable data received from the image processing device 12. In this example, the printing apparatus 14 performs color printing using each color of CMYK ink as a process color. The printing apparatus 14 may further perform printing using other colors of ink.

  FIG. 1B is a diagram illustrating an outline of quantization processing performed by the image processing apparatus 12. The image processing device 12 forms a pseudo halftone image that is a halftone image by quantizing the original image for each process color. This quantization is a process of converting the density value In (x, y) at each coordinate of the original image into a quantized value out (x, y) at the same coordinate of the pseudo halftone image.

  In this example, the image processing apparatus 12 performs this quantization by a method (hybrid error diffusion processing) that uses both dither matrix noise D (i, j) and cumulative error E (x, y). Further, the image processing device 12 sets the parameters related to the dither matrix noise D (i, j) and the accumulated error E (x, y) to values in the range of 0 to 1 (0% to 100%), respectively. The set noise usage rate Rn and error usage rate Re are further used.

  The noise usage rate Rn is a parameter indicating the degree of influence of the dither matrix noise D (i, j) on the quantization process, and is calculated according to the density value In (x, y) of the pixel to be quantized. . The image processing device 12 does not use the dither matrix noise D (i, j) as it is, but performs quantization using the product of the dither matrix noise D (i, j) and the noise usage rate Rn. Thereby, the image processing apparatus 12 uses the dither matrix noise D (i, j) according to the noise usage rate Rn corresponding to each pixel.

  The error usage rate Re is a parameter indicating the degree to which the accumulated error E (x, y) affects the quantization process, and is calculated according to the density value In (x, y) of the pixel to be quantized. The The image processing apparatus 12 does not use the accumulated error E (x, y) as it is, but uses the product of the accumulated error E (x, y) and the error usage rate Re, so that the density value In (x, y) of each pixel is used. An error-corrected input value In ′ (x, y) corresponding to y) is calculated. Then, quantization processing is performed using the calculated error-corrected input value In ′ (x, y). Accordingly, the image processing device 12 uses the accumulated error E (x, y) according to the error usage rate Re corresponding to each pixel. The quantization process will be described in detail later.

  According to this example, for example, the spatial frequency characteristic (dither characteristic) of the dither process can be influenced on the spatial frequency characteristic (error diffusion characteristic) of the error diffusion process. In addition, this enables, for example, quantization to be performed by a method in which a dither characteristic is incorporated into the error diffusion characteristic. Further, in this case, unlike the case of simply switching between dither processing and error diffusion processing, for example, it is possible to appropriately prevent the occurrence of a boundary line due to the switching of processing.

  Therefore, according to this example, for example, it is possible to perform a quantization process that appropriately utilizes a portion that each process is good at. In addition, for example, various problems relating to printing that occur in the quantization process can be appropriately solved, and the quantization process can be performed by a more appropriate method.

  The dither matrix noise D (i, j) is a value designated by a preset dither matrix, for example. The dither matrix noise D (i, j) may be the same as or similar to the dither matrix noise used in the conventional dither processing, for example. As the dither matrix noise D (i, j), for example, it is preferable to use noise having a blue noise characteristic. The image processing apparatus 12 preferably changes the dither matrix to be used for each process color.

  The accumulated error E (x, y) is a value obtained by accumulating the quantization error Q (x, y), which is an error generated in the quantization of the surrounding pixels, and uses a preset diffusion filter (diffusion matrix). Calculated. Further, the calculated accumulated error E (x, y) is stored in, for example, an error buffer. The image processing apparatus 12 calculates the accumulated error E (x, y) by the same or similar method as the accumulated error used in the conventional error diffusion process, for example.

  FIG. 2 is a flowchart illustrating an example of an operation for performing quantization. The image processing apparatus 12 performs the following processing for each process color, for example.

  In the operation of performing quantization, the image processing apparatus 12 of this example first selects a pixel to be quantized from the original image (pixel selection processing S102). Then, according to the density value In (x, y) of the selected pixel, a noise usage rate Rn and an error usage rate Re corresponding to the pixel are calculated. As a result, the image processing apparatus 12 determines the noise usage rate Rn and the error usage rate Re used for quantization of the pixel (usage rate determination processing S104).

  Subsequently, the image processing apparatus 12 calculates an error-corrected input value In ′ (x, y) that is a density value after correction by the accumulated error E (x, y) (error-corrected input value calculation). Process S106). In this process, for example, the image processing apparatus 12 calculates the product of the error usage rate Re corresponding to the pixel selected in the pixel selection process S102 and the accumulated error E (x, y), and the density value In ( x, y), and the value after the addition is calculated as an error-corrected input value In ′ (x, y).

  The image processing apparatus 12 further calculates a noise corrected threshold Th ′ that is a threshold reflecting the dither matrix noise D (i, j) as a threshold used in quantization (noise corrected threshold calculation processing). S108). In this process, for example, the image processing apparatus 12 calculates a product of the noise usage rate Rn corresponding to the pixel selected in the pixel selection process S102 and the dither matrix noise D (i, j) as a preset initial threshold value. The value after addition is calculated as the noise-corrected threshold Th ′.

  Then, the image processing device 12 compares the calculated noise corrected threshold value Th ′ with the error corrected input value In ′ (x, y). As a result, the image processing apparatus 12 performs quantization on the pixel selected in the pixel selection process S102 (quantization execution process S110).

  Subsequently, the image processing apparatus 12 diffuses the quantization error Q (x, y) generated by the quantization of the pixels to the peripheral pixels according to the diffusion filter (error distribution process S112). As a result, the image processing apparatus 12 adds the quantization error Q (x, y) to the value of the diffusion filter corresponding to the distribution destination coordinates of the peripheral pixels, and the accumulated error E (x, y) corresponding to each of the peripheral pixels. Update the value of y).

  Further, after the error distribution process S112, the image processing apparatus 12 determines whether or not the pixel on which the quantization has been performed is the final pixel of the original image (final pixel determination process S114). When it is determined that the pixel is the final pixel (S114: Yes), the quantization process for the original image is terminated. If it is determined that the pixel is not the last pixel (S114: No), the process proceeds again to the pixel selection process S102, and the next pixel is selected. Thereby, the image processing apparatus 12 sequentially selects pixels to be quantized in the pixel selection process S102. In addition, by performing the processing subsequent to the pixel selection processing S102 for the sequentially selected pixels, quantization for the pixels is executed.

  According to this example, for example, by adding the accumulated error E (x, y) multiplied by the error usage rate Re to the density value In (x, y), the degree of influence of the error diffusion characteristic can be appropriately adjusted. . Further, by adding the dither matrix noise D (i, j) multiplied by the noise usage rate Rn to the initial threshold value Th, the degree of influence of the dither characteristic can be adjusted appropriately.

  Thereby, for example, the degree of influence of each of the error diffusion characteristic and the dither characteristic can be appropriately set according to the density value In (x, y) that is the input value. Further, it is possible to perform a quantization process that appropriately utilizes a portion that each of the error diffusion process and the dither process is good at. Hereinafter, each process during the operation of performing quantization will be described in more detail.

  FIG. 3 is a diagram for explaining the pixel selection process S102 in more detail. FIG. 3A shows an example of the order in which pixels are sequentially selected. In the pixel selection process S102 of this example, for example, the image processing apparatus 12 sequentially selects pixel lines, and sequentially selects pixels in the selected lines along a predetermined processing direction. Further, the image processing apparatus 12 further switches the processing direction for each line to be processed. For example, the image processing apparatus 12 sequentially selects pixels from left to right in odd lines and from right to left in even lines. As a result, the image processing apparatus 12 performs the quantization process by bidirectional processing.

  FIG. 3B is a diagram illustrating an example of the effect of the bidirectional processing. When the unidirectional processing is performed in which the processing direction is only one direction, and when the bidirectional processing is performed, An example of the result of quantization is shown. When bidirectional processing is performed, the error diffusion direction is not constant, so that dots can be more appropriately dispersed. As a result, the generation of worm noise or the like can be more appropriately prevented as compared with the case where quantization is performed by, for example, unidirectional processing.

  4 to 6 are diagrams for explaining the usage rate determination process S104 in more detail. FIG. 4 shows an example of a graph and a calculation formula for associating the error usage rate Re and the noise usage rate Rn with the density value In (x, y) (hereinafter referred to as the input value In).

  In this example, the image processing apparatus 12 uses the error usage rate Re and the noise usage based on a function that continuously changes with respect to the input value In in the range of 0 or more which is the minimum input value MinIn and less than or equal to the maximum input value MaxIn. The rate Rn is calculated. The maximum input value MaxIn and the minimum input value MinIn are, for example, the maximum value and the minimum value in the range that can be taken by the density value that becomes the input value In.

  In this function, the image processing apparatus 12 uses the highlight-side error usage minimum density value Hes, which is an example of the third highlight reference value, as the reference indicating the density range in the highlight portion, and the first highlight. A highlight side noise usage maximum density value Hn, which is an example of a reference value, and a highlight side error usage ratio maximum density value He, which is an example of a second highlight reference value, are used. In addition, as a reference indicating the density range in the shadow portion, the shadow-side error usage maximum density value Se that is an example of the first shadow reference value and the shadow-side noise usage maximum maximum density value that is an example of the second shadow reference value. Sn and the shadow-side error usage minimum density value Ses, which is an example of the third shadow reference value, are used. Further, as a reference indicating the density range sandwiching the initial threshold Th at the center of the halftone portion, the highlight side noise usage rate 0% density value Hnz, the shadow side noise usage rate 0% density value Snz, and the first halftone reference value C1 , And the second halftone reference value C2.

  These parameters are set so that at least Hes ≦ Hn <He <Se <Sn ≦ Ses and Hn <Hnz <Snz <Sn. In this example, these parameters are 0 (MinIn) <Hes ≦ Hn <He <C1 <Hnz <Th <Snz <C2 <Se <Sn ≦ Ses <MaxIn, according to the magnitude relationship shown in the graph. Is set as follows.

  In the usage rate determination process S104, the image processing apparatus 12 determines the error usage rate Re and the noise usage rate Rn according to the calculation formula shown below the graph. However, when the noise usage rate Rn calculated by this mathematical formula is smaller than the predetermined minimum noise usage rate RnMin, the image processing apparatus 12 sets the noise usage rate Rn to the minimum noise usage rate RnMin. Accordingly, the image processing apparatus 12 sets the noise usage rate Rn to a value equal to or higher than the minimum noise usage rate RnMin regardless of the density value of the pixel.

  The error usage rate Re and the noise usage rate Rn are set to values within the range of 0 to 100% (values 0 to 1). In the calculation formula shown below the graph, when it is 100% or more, it is set to 100%. If it is 0% or less, it is set to 0%.

  Further, the minimum noise usage rate RnMin is preset to a value larger than 0, for example, at the time of adjustment at the time of parameter setting. The minimum noise usage rate RnMin may be set to a value of 0.1 (10%) or more, for example. For example, the minimum noise usage rate RnMin is preferably 0.1 to 0.2 (10 to 20%), for example. As can be seen from the graph, when the input value In is equal to the first halftone reference value C1 or the second halftone reference value C2, the noise usage rate Rn calculated by the calculation formula is the minimum noise usage rate RnMin. Is equal to

  With the above method, for example, when the input value In is equal to or higher than the highlight-side error usage rate maximum density value He and equal to or lower than the shadow-side error usage rate maximum density value Se, the image processing apparatus 12 sets the error usage rate Re to Set to 1 (100%). Further, for example, when the input value In is equal to or less than the highlight-side error usage rate minimum density value Hes, the error usage rate Re is set to zero. When the input value In is not less than the highlight side error usage rate minimum density value Hes and not more than the highlight side error usage rate maximum density value He, the error usage rate Re is (In-Hes) / (He-Hes). Set to the calculated value. Thus, for example, when the input value In is less than or equal to the highlight side error usage rate maximum density value He, the error usage rate Re is a value between 0 and 1 (100%), and the highlight side error usage rate maximum density. It is set to a value gradually decreased from 1 according to the difference between the value He and the input value In.

  For example, when the input value In is not less than the shadow-side error use rate maximum density value Se and not more than the shadow-side error use rate minimum density value Ses, the error use rate Re is set to (Ses-In) / (Ses-Se). ) To be calculated. When the input value In is greater than or equal to the shadow side error use rate minimum density value Ses, the error use rate Re is set to zero. Thus, for example, when the input value In is equal to or greater than the shadow-side error usage rate maximum density value Se, the error usage rate Re is a value between 0 and 1 (100%), and the input value In and the shadow-side error usage are It is set to a value gradually decreased from 1 in accordance with the difference from the maximum rate density value Se.

  In this case, the error usage rate Re from the highlight portion to the halftone portion gradually increases from the highlight side error usage rate minimum density value Hes with respect to the input value In, for example, and the highlight side error usage rate maximum density. The value He is the maximum value. Further, the error usage rate Re from the shadow portion to the halftone portion gradually decreases from the shadow side error usage rate maximum density value Se with respect to the input value In, and becomes the minimum value in the shadow side error usage rate minimum density value Ses. Become.

  Thereby, for example, the image processing apparatus 12 uses the error usage rate Re when the input value In is a density value corresponding to either the highlight part or the shadow part as the density value corresponding to the halftone part. Is set to a value smaller than the error usage rate Re. In this case, for example, the error usage rate Re has a value of 0 (0%) at both ends of the density range, and is in front of the texture generation units Hd and Sd, which are density ranges in which a texture peculiar to dither processing is generated. It changes so that it may become value 1 (100%). If comprised in this way, the structure which mainly uses the error usage rate Re by a halftone can be implement | achieved appropriately, for example.

  For example, when the input value In is less than or equal to the highlight side noise usage rate maximum density value Hn, or when the input value In is greater than or equal to the shadow side noise usage rate maximum density value Sn, the image processing apparatus 12 sets the noise usage rate Rn to 1 Set to (100%). When the input value In is equal to or higher than the first halftone reference value C1 and equal to or lower than the second halftone reference value C2, the noise usage rate Rn is set to the lowest noise usage rate RnMin.

  For example, when the input value In is not less than the highlight side noise use rate maximum density value Hn and not more than the first halftone reference value C1, the image processing apparatus 12 sets the noise use rate Rn to be not less than the minimum noise use rate RnMin. In addition, the value is 1 (100%) or less, and is set to a value that is gradually decreased from 1 in accordance with the difference between the input value In and the highlight side noise use rate maximum density value Hn. Further, for example, when the input value In is equal to or greater than the second halftone reference value C2 and equal to or less than the shadow side noise usage rate maximum density value Sn, the image processing apparatus 12 sets the noise usage rate Rn to be equal to or greater than the minimum noise usage rate RnMin. And a value of 1 (100%) or less, and is set to a value that is gradually increased from the minimum noise usage rate RnMin according to the difference between the input value In and the second halftone reference value C2.

  Thereby, for example, the image processing apparatus 12 uses the noise usage rate Rn when the input value In is a density value corresponding to either the highlight part or the shadow part as the density value corresponding to the halftone part. The noise usage rate Rn is set to a larger value. In this case, for example, the noise usage rate Rn has a value of 1 (100%) in the highlight portion and the shadow portion, and changes so that the value gradually decreases as it becomes a halftone.

  According to this example, for example, the noise usage rate Rn is increased and the error usage rate Re is decreased in a highlight portion and a shadow portion where dot delay is likely to occur when the influence of the error diffusion characteristic is strong. Thereby, for example, in the highlight portion and the shadow portion, the influence of the dither characteristics that can disperse and arrange the dots becomes large, so that the occurrence of dot delay can be appropriately suppressed.

  Also, by increasing the influence of the error diffusion characteristic and performing the quantization process in the halftone part, for example, a more natural pseudo gradation can be obtained. Further, for example, in the density range that becomes the texture generation unit, the dot usage can be changed by lowering the noise usage rate Rn and increasing the error usage rate Re. Thereby, for example, generation of texture can be appropriately suppressed.

  Further, for example, by providing the minimum noise usage rate RnMin so that the noise usage rate Rn does not become 0, for example, in the halftone part, the influence of the error diffusion characteristic is slightly affected while slightly affecting the dither characteristic. Can be big. Thereby, for example, pattern noise can be appropriately suppressed.

  Further, for example, the dither characteristic is dominant by gradually changing the magnitude of the influence of the dither characteristic and the magnitude of the influence of the error diffusion characteristic according to the calculation formulas for calculating the error usage rate Re and the noise usage rate Rn. It is possible to appropriately suppress the occurrence of the boundary line at the switching portion between the region and the region where the error diffusion characteristic is dominant. In addition, this enables smooth switching of the quantization processing method.

  Therefore, according to this example, each of the error usage rate Re and the noise usage rate Rn can be appropriately set according to, for example, the density value of the pixel. This also makes it possible to appropriately perform a quantization process that makes better use of the parts that the error diffusion process and the dither process are good at.

  Next, the process of calculating the error usage rate Re and the noise usage rate Rn will be described in more detail. FIG. 5 is a flowchart illustrating an example of processing for calculating the error usage rate Re.

  In calculating the error usage rate Re, the image processing apparatus 12 first determines whether or not the input value In is in a range not less than the highlight side error usage rate maximum density value He and not more than the shadow side error usage rate maximum density value Se. Is determined (S202). And when it determines with it being in this range (S202: Yes), the error usage rate Re is set to 1 (100%) which is the maximum usage rate (S204).

  Further, when it is determined that it is not within this range (S202: No), the input value In is larger than the highlight side error usage rate minimum density value Hes and is higher than the highlight side error usage rate maximum density value He. It is determined whether it is in a small range (S206). And when it determines with it being in this range (S206: Yes), the error usage rate Re is set to the value calculated by Re = (In-Hes) / (He-Hes) (S208).

  If it is determined in S206 that the input value In is not within the range (S206: No), the image processing apparatus 12 further determines that the input value In is greater than the shadow-side error usage maximum density value Se, and It is determined whether or not it is in a range smaller than the shadow-side error usage rate minimum density value Ses (S210). And when it determines with it being in this range (S210: Yes), the error usage rate Re is set to the value calculated by Re = (Ses-In) / (Ses-Se) (S212). If it is determined in S210 that the input value In is not within the range (S210: No), the image processing apparatus 12 sets the error usage rate Re to 0 (0%) (S214).

  Then, the image processing apparatus 12 employs the error usage rate Re set in any of S204, S208, S212, or S214 as the error usage rate Re corresponding to the input value In (S216). According to this example, the error usage rate Re can be calculated appropriately.

  FIG. 6 is a flowchart illustrating an example of processing for calculating the noise usage rate Rn. In calculating the noise usage rate Rn, the image processing apparatus 12 first sets the input value In to a range that is larger than the highlight side noise usage rate maximum density value Hn and smaller than the highlight side noise usage rate 0% density value Hnz. It is determined whether or not there is (S302). And when it determines with it being in this range (S302: Yes), the noise usage rate Rn is set to the value calculated by Rn = (Hnz-In) / (Hnz-Hn) (S304).

  Further, when it is determined that it is not within this range (S302: No), the input value In is further larger than the shadow side noise usage rate 0% density value Snz and smaller than the shadow side noise usage rate maximum density value Sn. It is determined whether or not there is (S306). And when it determines with it being in this range (S306: Yes), the noise usage rate Rn is set to the value calculated by Rn = (In-Snz) / (Sn-Snz) (S308).

  Then, after setting the noise usage rate Rn to the calculated value in S304 or S308, it is determined whether or not the set noise usage rate Rn is larger than the minimum noise usage rate RnMin which is the minimum usage rate (S310). If the noise usage rate Rn is equal to or lower than the lowest noise usage rate RnMin (S310: No), the value of the noise usage rate Rn is changed to the lowest noise usage rate RnMin (S312), and the process proceeds to S316. When the noise usage rate Rn is greater than the minimum noise usage rate RnMin (S310: Yes), the process proceeds to S316 without changing the value of the noise usage rate Rn.

  If it is determined in S306 that the input value In is not within the above range (S306: No), the image processing apparatus 12 sets the noise usage rate Rn to 1 (100%) which is the maximum usage rate. (S314). Then, the image processing apparatus 12 employs the noise usage rate Rn set in any of S304, S308, S312 or S314 as the noise usage rate Rn corresponding to the input value In (S316). According to this example, the noise usage rate Rn can be calculated appropriately.

  FIG. 7 is a flowchart illustrating an example of the operation of the quantization execution process S110. In the quantization execution process S110 of this example, the image processing apparatus 12 first determines whether or not the input value In and the maximum input value MaxIn are equal (maximum value determination process S402). If it is determined that they are equal (S402: Yes), the output value indicating the quantization result is set to 1 (S404). This value 1 is an example of a value to be output when the input value In is larger than the threshold value Th. In addition to setting the output value, the error value used in the subsequent error distribution process S112 is set to In'-MaxIn, which is the difference between the error-corrected input value In 'and the maximum input value MaxIn.

  When it is determined in S402 that the input value In and the maximum input value MaxIn are not equal (S402: No), the image processing apparatus 12 further determines whether or not the input value In and the minimum input value MinIn are equal. Determination (minimum value determination processing S406). If it is determined that they are equal (S406: Yes), the output value is set to 0 (S408). This value 0 is an example of a value to be output when the input value In is smaller than the threshold value Th. In addition to setting the output value, the error value is set to the error-corrected input value In ′.

  If it is determined in S406 that the input value In and the minimum input value MinIn are not equal (S406: No), the image processing apparatus 12 has the error-corrected input value In ′ larger than the noise-corrected threshold value Th ′. It is determined whether or not (S410). If it is determined that the value is large (S410: Yes), the output value is set to 1 and the error value is set to In'-MaxIn (S412). If it is determined in S410 that the error-corrected input value In ′ is equal to or less than the noise-corrected threshold Th ′ (S410: No), the output value is set to 0 and the error value is set to In ′ (S414). .

  Then, the image processing apparatus 12 acquires the output value and error value set in S404, S408, S412, or S414 as the quantized value and error value resulting from the quantization execution process S110 (quantized value acquisition process). S416). Further, the acquired error value is transferred to a subsequent error distribution process S112. According to this example, for example, the output value and the error value can be easily and appropriately set.

  8 and 9 are diagrams for explaining the error distribution processing S112 in more detail. FIG. 8 shows an example of the diffusion filter used in this example. FIG. 8A shows an example of a diffusion filter used in each direction of bidirectional processing.

  In this example, the diffusion filter is, for example, a matrix of Jarvis, Judice & Ninke. Also, the illustrated filters are used for each of the main scanning direction and the reverse main scanning direction, which are the directions of bidirectional processing. In the illustrated filter, the position where the symbol * is entered is the coordinates [0, 0] (origin) of the input value In. The numerical value in each matrix is a distribution ratio when an error is distributed to surrounding pixels.

  FIG. 8B shows an example of how to distribute errors. In the process of distributing an error to surrounding pixels, when the error distribution destination coordinates are out of the image width, the image processing apparatus 12 changes the distribution destination coordinates to the first coordinates of the next line. The same processing is performed when the opposite coordinate is out of range. Further, when there is no distribution destination line, the image processing apparatus 12 does not distribute the error. Also, no error is distributed when the coordinates of the distribution destination are processed pixels.

  FIG. 9 is a flowchart showing an example of the operation of the error distribution process S112. In the error distribution process S112, the image processing apparatus 12 executes a loop for sequentially changing the Y coordinate between steps S502 and S526 in the flowchart. In this loop, the image processing apparatus 12 increases the value of Y by 1 between 0 and the diffusion matrix height. The diffusion matrix height is, for example, the number of matrix rows used as a diffusion filter.

  Further, a loop for sequentially changing the X coordinate is executed between steps S504 and S524. In this loop, the image processing apparatus 12 increases the value of X by 1 between 0 and the diffusion matrix width. The diffusion matrix width is, for example, the number of matrix columns used as a diffusion filter.

  In these loops, the image processing apparatus 12 first sets the distribution destination coordinates (X ′, Y ′) as shown in the flowchart. Further, a distribution ratio is set according to the diffusion filter (S506). First, at this time, it is determined whether or not the distribution ratio is greater than 0 (S508). If the distribution ratio is 0 or less (S508: No), the process proceeds to S504 again without distributing the error.

  If the distribution ratio is greater than 0 (S508: Yes), the process proceeds to S510 and subsequent steps. In this process, if the distribution destination coordinate X ′ is smaller than the image width (S510: Yes), the coordinate X ′ is 0 or more (S512: Yes), and the distribution destination coordinate Y ′ is smaller than the image height (S514: Yes), the error value to be distributed (quantization error) is set to the product of the error value and the distribution ratio (S516), and added to the distribution error value stored in the cumulative error buffer (S518). Thereby, the image processing apparatus 12 updates the accumulated error in accordance with the generated quantization error.

  In S510, when the coordinate X 'is equal to or larger than the image width (S510: No), the coordinate X' = X'-image width and the coordinate Y '= Y' + 1 are set (S520), and the process proceeds to S514. If the coordinate X ′ is smaller than 0 in S512 (S512: No), the coordinate X ′ = X ′ + image width and the coordinate Y ′ = Y ′ + 1 are set (S522), and the process proceeds to S514. In S514, if the coordinate Y 'is equal to or higher than the image height (S514: No), the process proceeds to S504 without distributing the error.

  With the above operation, the image processing apparatus 12 adds the quantization error to the value stored in the cumulative error buffer and calculates the cumulative error. According to this example, the error can be appropriately distributed. This also makes it possible to calculate the accumulated error appropriately.

  FIG. 10 shows an example of the print result according to this example together with the print result according to the reference example. FIGS. 10A to 10C show printing results according to a reference example different from the present example described with reference to FIGS. FIG. 10A shows an example of a print result when only a conventionally known typical error diffusion process is performed. FIG. 10B shows an example of a print result when only a conventionally known dither process is performed.

  When only typical error diffusion processing is performed, for example, a problem such as dot delay occurs in a highlight portion or a shadow portion. Further, when only the dither process is performed, for example, a problem such as a texture peculiar to the dither process occurs.

  FIG. 10C shows an example of a print result when the dither process and the error diffusion process are simply switched depending on the density range. In this example, for example, switching processing for gradually changing the noise usage rate and the error usage rate is performed between the highlight portion and the halftone portion or between the halftone portion and the shadow portion as in this example. Instead, dither processing and error diffusion processing are switched. As described above, when the dithering process and the error diffusion process are simply switched without performing an appropriate switching process, there is a problem that a boundary line is generated due to the switching.

  FIG. 10D shows an example of the print result according to this example. As can be seen from the figure, according to this example, it is possible to appropriately suppress, for example, dot delay in highlight portions and shadow portions. In addition, it is possible to appropriately prevent the occurrence of texture in halftones and the generation of boundary lines due to switching. Therefore, according to this example, the quantization process can be performed more appropriately.

  Here, the same or similar quantization processing as described above can be applied to multi-gradation quantization. Accordingly, multi-tone quantization processing will be described below.

  The multi-tone quantization process is, for example, a process of converting a density value indicating the color density of each pixel in an image into a discrete value of three or more stages. The three or more discrete values are, for example, the lowest quantized value corresponding to the result of quantizing the minimum density value and two or more discrete values larger than the lowest quantized value. And a quantization result value of at least. The quantization result values of two or more levels are values corresponding to the result of quantizing a value larger than the minimum density value by a predetermined amount or more, for example.

  Multi-tone quantization can be performed by the same or similar configuration as that shown in FIG. In this case, the image processing apparatus 12 performs multi-tone quantization processing based on a preset program during the image forming processing. Further, the printing device 14 performs multi-tone printing by ejecting a plurality of types of ink droplets having different dot sizes. In this case, of the quantization results, the lowest quantized value is associated with a state where no dots are formed. The quantization result value at each stage is associated with each dot size.

  FIG. 11 is a flowchart showing a first example of multi-tone quantization processing according to an embodiment of the present invention. The printing apparatus 14 performs the following processing according to a preset program.

  In the configuration described below, the printing apparatus 14 ejects ink droplets corresponding to dots of three types of sizes, large, medium, and small. In this case, the values that can be obtained as a result of quantization are the sum including the lowest quantization value corresponding to the state in which no dots are formed and the three-stage quantization result values corresponding to the large, medium, and small dot sizes, respectively. There are four levels of discrete values. Further, the discrete values in four stages are associated with, for example, 2-bit numerical values 00, 01, 10, and 11.

  Further, the printing apparatus 14 ejects a plurality of types of ink droplets having different dot sizes by changing the size of the droplets ejected from the nozzles of the inkjet head, for example. The printing apparatus 14 may eject a plurality of types of ink droplets having different dot sizes by changing the number of ink droplets to be landed in one place on the medium.

  In this flowchart, the image processing apparatus 12 executes a loop for sequentially changing the Y coordinate between steps S602 and S616. In this loop, the image processing apparatus 12 increases the value of Y by 1 between 0 and the image height. Further, a loop for sequentially changing the X coordinate is executed between steps S604 and S614. In this loop, the image processing device 12 increases the value of X by 1 between 0 and the image width. Note that the execution of this loop in this flowchart is the same as or similar to the final pixel determination process S114 in the flowchart shown in FIG.

  Then, the image processing apparatus 12 sequentially selects pixels to be quantized (pixel selection process S606). This step is, for example, the same or similar processing as the pixel selection processing S102 in the flowchart shown in FIG. For example, the image processing apparatus 12 sequentially selects pixel lines in accordance with the above loop, and sequentially selects pixels in the selected line along a predetermined processing direction. Further, the image processing apparatus 12 further switches the processing direction for each line to be processed. For example, the image processing apparatus 12 sequentially selects pixels from left to right in odd lines and from right to left in even lines. As a result, the image processing apparatus 12 performs the quantization process by bidirectional processing.

  Subsequently, the image processing apparatus 12 is a value that is used as an input value corresponding to each dot size in the subsequent processing based on the pixel density value that is the input value for the entire multi-tone quantization process. Each input value is calculated (input value calculation processing S608). In this process, the image processing apparatus 12 sets the input value for each stage corresponding to each dot size according to the density value of the pixels sequentially selected in the pixel selection process S606 based on a preset dot size determination table. calculate.

  The dot size determination table will be described in detail later. In the following description, an input value for each stage corresponding to a large dot size (hereinafter referred to as a large dot) is an L curve value. An input value for each stage corresponding to a medium dot size (hereinafter referred to as “Middle dot”) is an M curve value. An input value for each stage corresponding to a small dot size (hereinafter referred to as a Small dot) is an S curve value.

  Further, in the input value calculation process S608 of this example, the image processing apparatus 12 further sets an output value (output dot size) that is a result of multi-gradation quantization for the pixel selected in the pixel selection process S606. , It is initialized to 0, which is a value corresponding to a state where no dots are formed.

  Subsequently, the image processing apparatus 12 calculates an error-corrected input value and a noise-corrected threshold value (input value threshold value correcting process S610). In this example, the image processing apparatus 12 performs the usage rate determination process S104, the error-corrected input value calculation process S106, and the noise-corrected threshold value calculation described with reference to FIG. 2 for the input value for each stage corresponding to each dot size. The same or similar processing as the processing S108 is performed. As a result, the image processing apparatus 12 calculates different error-corrected input values and noise-corrected threshold values for each dot size.

  Subsequently, the image processing device 12 performs quantization (quantization execution processing S612). In the quantization execution process S612 of this example, the image processing apparatus 12 performs, for example, a process performed for each dot size using the input value for each stage, the error-corrected input value, and the noise-corrected threshold value corresponding to each dot size. A step-by-step determination process is performed. Based on those results, the result of multi-tone quantization is determined. Accordingly, for example, the image processing apparatus 12 uses dither matrix noise according to the noise usage rate corresponding to the pixel for the density value of the pixel sequentially selected in the pixel selection processing S606, and applies the pixel to the pixel. Multi-tone quantization is performed using the accumulated error according to the corresponding error usage rate.

  In this example, the image processing apparatus 12 performs an error distribution process in the quantization execution process S612. The image processing apparatus 12 may execute the error distribution process in another step separately from, for example, the quantization execution process S612.

  After the quantization execution process S612, the image processing apparatus 12 proceeds to step S614 and proceeds with the next loop. According to this example, multi-tone quantization processing can be performed appropriately.

  Next, each process described with reference to FIG. 11 will be described in more detail. FIG. 12 is a diagram for explaining the input value calculation process S608 and the input value threshold value correction process S610 in more detail. FIG. 12A is a graph showing an example of a dot size determination table used in the input value calculation process S608.

  The dot size determination table is data indicating the ratio of the dot size to be output for each density. For example, the ratio of the dot size to the number of dots per unit area at a certain density is determined. In this case, the total ratio of the respective dot sizes is 0 to 100%. The dot size determination table is an example of input value calculation data for each stage. The input value calculation data for each stage is data that determines, for each dot size, the number of dots of ink droplets to be formed per unit area according to the density value of the pixel.

  In this example, as shown in the graph, the dot size determination table defines input values for each stage corresponding to each of a large dot, a middle dot, and a small dot with respect to the density value of the pixel. Based on such a dot size determination table, in the input value calculation process S608, the image processing apparatus 12 inputs, for each pixel size value sequentially selected in the pixel selection process S606, an input value corresponding to each dot size. Is calculated according to the number of dots associated with the density value for the dot size in the dot size determination table.

  In the dot size determination table of this example, the M curve value and the L curve value are 0 below a predetermined density N1. In addition, the L curve value is 0 at a predetermined density N2 or less greater than N1.

  The S curve value gradually increases with a predetermined slope m1 in the range between the density value 0 and N1 or the vicinity thereof, and gradually increases with the slope m2 smaller than m1 in the range where the density value becomes more than that. To do. The M curve value gradually increases with a predetermined slope m3 in the range between the density value N1 and N2 or the vicinity thereof, and gradually increases with a slope m4 smaller than m3 in the range where the density value is more than that. To do. The L curve value gradually increases with a predetermined slope m5 in a range between the density value N2 and a predetermined density N3 larger than N2, and is larger than m5 in a range where the density value becomes more than that. It gradually increases with an inclination m6.

  Further, the S curve value and the M curve value intersect at a density value C1 between N1 and N2. The S curve value and the L curve value intersect at a density value C2 greater than N3. Further, the M curve value and the L curve value intersect at a density value C3 larger than C2. As a result, in the range where the density value is less than C1, S curve value> M curve value> L curve value. When the density value is in the range from C1 to C2, M curve value> S curve value> L curve value. When the density value is in the range of C2 to C3, M curve value> L curve value> S curve value. In the range where the density value exceeds C3, L curve value> M curve value> S curve value.

  FIG. 12B is a diagram for explaining the input value threshold value correction processing S610, and shows a graph associating the error usage rate Re, the noise usage rate Rn, and the input value for each stage. In this example, the image processing apparatus 12 responds to each of the L-curve value, the M-curve value, and the S-curve value, which are input values for each stage calculated in the input value calculating process S608, based on the relationship shown in the graph. An error usage rate Re and a noise usage rate Rn are calculated. Then, the image processing device 12 calculates an error-corrected input value and a noise-corrected threshold value for each dot size based on the calculated error usage rate Re and noise usage rate Rn.

  Note that the graph shown in FIG. 12B is different from the graph shown in FIG. 4 in that, for example, the input value at each stage is used instead of the input value In, and the display orientation is changed in accordance with FIG. It is a thing. With this configuration, for example, the error usage rate Re and the noise usage rate Rn corresponding to each dot size can be calculated appropriately.

  In this example, the image processing apparatus 12 calculates the error usage rate Re and the noise usage rate Rn based on the same graph for all dot sizes, for example. In this case, for example, the same parameters as in FIG. 4 can be used as the parameters (Hes, Hn,...) In the graph. Each parameter may be changed for each dot size.

  FIG. 13 is a flowchart illustrating an example of more specific processing in the input value threshold value correction processing S610. In the input value threshold value correction process S610 of this example, the image processing apparatus 12 first, for example, an L curve value, an M curve value, and an S curve value that are input values for each stage calculated in the input value calculation process S608, and Based on the relational expression shown in the graph of FIG. 12B, the noise L curve value, M curve value, S curve value usage rate, and error usage rate corresponding to each dot size are calculated (noise usage rate calculation processing). S620, error usage rate calculation processing S622). In these processes, the image processing apparatus 12 uses, for example, the L curve value, the M curve value, and the S curve value instead of the input value In in the process described with reference to FIGS. The noise usage rate and error usage rate for each dot size are calculated.

  Then, by correcting each of the L curve value, the M curve value, and the S curve value based on the calculated error usage rate, an error corrected input value corresponding to each dot size is calculated (error corrected input value). Calculation process S624). In addition, a noise-corrected threshold value is calculated by correcting a preset threshold value based on the calculated noise usage rate (noise-corrected threshold value calculation process S626).

  In the error-corrected input value calculation process S624 of this example, the image processing apparatus 12 calculates an error-corrected input value using the accumulated error prepared for each dot size. In the error-corrected input value calculation process S624 and the noise-corrected threshold value calculation process S626, the image processing apparatus 12 replaces the input value In with the L curve value in the process described with reference to FIG. By using each of the M curve value and the S curve value, an error corrected input value and a noise corrected threshold value corresponding to each dot size are calculated. As a result, the image processing apparatus 12 acquires the error-corrected input value and the noise-corrected threshold value for each dot size, and proceeds to the next quantization execution process S612.

  FIG. 14 is a flowchart illustrating an example of the process in the quantization execution process S612. In this example, the image processing apparatus 12 performs a Large process S640 that is a quantization and error distribution process corresponding to a Large dot, a Middle process S650 that is a quantization and error distribution process corresponding to a Middle dot, and a Small process. Small processing S660, which is quantization and error distribution processing corresponding to dots, is performed in this order.

  As a result, the image processing apparatus 12 gives priority to the processing corresponding to the dot size larger than the quantization processing priority corresponding to each dot size. More specifically, for example, as in this example, when three types of dot sizes, Large dots, Middle dots, and Small dots, are set, the image processing apparatus 12 performs quantization corresponding to the Large dots. Multi-tone quantization processing is performed by setting the processing priority to the highest and the quantization processing priority corresponding to the Small dot to the lowest.

  Next, the contents of the Large quantization process S642, the Middle quantization process S652, and the Small quantization process S662 will be described. In the large process S640 in this example, the image processing apparatus 12 performs a large quantization process S642 and a large error distribution process S644. The Large quantization process S642 is an example of a stage-by-stage determination process corresponding to a Large dot. In the Large quantization process S642, for example, the image processing apparatus 12 performs a two-tone quantization process using the error-corrected input value and the noise-corrected threshold value corresponding to the Large dot. As a result, the image processing apparatus 12 performs a two-gradation quantization process on the L curve value that is the input value for each stage corresponding to the Large dot. In this example, the image processing apparatus 12 further sets an error value caused by quantization in the large quantization process S642.

  Large error distribution processing S644 is an example of cumulative error calculation processing for each stage corresponding to large dots. In the Large error distribution process S644, for example, the image processing apparatus 12 distributes the error value set in the Large quantization process S642 to surrounding pixels as a dedicated error for Large dots. As a result, the image processing apparatus 12 updates the accumulated error in accordance with the result of the Large quantization process S642, and calculates a new accumulated error corresponding to the Large dot.

  In the middle processing S650, the image processing apparatus 12 performs the middle quantization processing S652 and the middle error distribution processing S654 in the same manner as the large processing S640. The middle quantization process S652 and the middle error distribution process S654 use the large quantization process S642 and the large error except that each parameter corresponding to the middle dot is used instead of each parameter corresponding to the large dot. This is the same process as the distribution process S644. In this case, the Middle quantization process S652 and the Middle error distribution process S654 are an example of a stage determination process and a stage cumulative error calculation process corresponding to a Middle dot.

  In the Small process S660, the image processing apparatus 12 performs the Small quantization process S662 and the Small error distribution process S664 in the same manner as the Large process S640. The Small quantization process S662 and the Small error distribution process S664 use the Large quantization process S642 and the Large error except that each parameter corresponding to the Small dot is used instead of each parameter corresponding to the Large dot. This is the same process as the distribution process S644. In this case, the Small quantization process S662 and the Small error distribution process S664 are an example of the determination process for each stage and the accumulated error calculation process for each stage corresponding to the Small dots.

  In this example, the image processing apparatus 12 is an accumulation error buffer provided for each dot size as an accumulation error buffer used in each of the error distribution process for large S644, the error distribution process for middle S654, and the error distribution process for small S664. Use error buffer. For example, in the large error distribution process S644, the image processing apparatus 12 accumulates the error generated in the large quantization process S642 in the large dot accumulation error buffer. The same applies to middle dots and small dots.

  Next, the Large quantization process S642 will be described in more detail. FIG. 15 is a flowchart illustrating an example of processing in the Large quantization processing S642. In the Large quantization process S642 in this example, the image processing apparatus 12 first sets the output value resulting from the multi-tone quantization to a value other than 0, which is a value corresponding to the case where no dot is formed. It is determined whether or not it has been performed (S702). If the output value is set to a value other than 0 (S702: Yes), only the error value is set without changing the output value (S704). In step S704, the image processing apparatus 12 sets the error value to the error-corrected input value In ′. Then, after setting the error value, the process proceeds to step S720.

  In this example, the output value to be determined in S702 is initialized to 0 in the input value calculation process S608 (see FIG. 11) before the quantization execution process S612 (see FIG. 11). For this reason, the output value becomes 0 at least at the start of the Large quantization process S642, which is executed first, among the Large quantization process S642, the Middle quantization process S652, and the Small quantization process S662. Yes. Thereby, the image processing apparatus 12 determines that the output value is set to 0 at least in Step S702 of the Large quantization process S642, and performs the two-tone quantization process described below.

  When the output value is set to 0 (S702: No), the image processing apparatus 12 further determines whether or not the input value In and the maximum input value MaxIn are equal (maximum value determination processing S706). If it is determined that they are equal (S706: Yes), the output value is set to a value corresponding to the Large dot (S708). In addition to setting the output value, the error value is set to In′−MaxIn, which is the difference between the error-corrected input value In ′ and the maximum input value MaxIn. Then, after setting these values, the process proceeds to step S720.

  When it is determined in S706 that the input value In and the maximum input value MaxIn are not equal (S706: No), the image processing apparatus 12 further sets the input value In and 0, which is the value of the minimum input value MinIn. Are equal to each other (minimum value determination processing S710). If it is determined that they are equal (S710: Yes), the output value is set to 0 (S712). In addition to setting the output value, the error value is set to the error-corrected input value In ′. Then, after setting these values, the process proceeds to step S720.

  If it is determined in S710 that the input value In is not equal to 0 (S710: No), the image processing apparatus 12 determines whether the error-corrected input value In ′ is greater than the noise-corrected threshold Th ′. Is determined (S714). And when it determines with it being large (S714: Yes), an output value is set to the value corresponding to a Large dot. Further, the error value is set to In'-MaxIn (S716). If it is determined in S714 that the error-corrected input value In ′ is equal to or less than the noise-corrected threshold Th ′ (S714: No), the output value is set to 0 and the error value is set to In ′ (S716). . Then, after setting these values, the process proceeds to step S720.

  After the above processing, the image processing apparatus 12 acquires an output value and an error value set after the above processing as a result of the Large quantization processing S642 (S720). This step S720 is an example of a quantization value acquisition process. Then, these values are transferred to the subsequent processing. In the subsequent processing, for example, in Large error distribution processing S644, error distribution is performed based on the error value. According to this example, the Large quantization process S642 can be appropriately performed. Further, in the large error distribution process S644, it is possible to appropriately distribute the error generated by the quantization to surrounding pixels.

  Here, in this example, the image processing apparatus 12 also performs processing according to the flowchart shown in FIG. 15 for the middle quantization processing S652 and the small quantization processing S662. For example, in the middle quantization process S652, the image processing apparatus 12 performs the same process as described above except that the parameter corresponding to the middle dot is used instead of the parameter corresponding to the large dot. In this case, in S708 and S716 of Middle quantization processing S652, the image processing apparatus 12 sets the output value to a value corresponding to the Middle dot. In addition, in the Small quantization process S662, the image processing apparatus 12 performs the same process as described above except that the parameter corresponding to the Small dot is used instead of the parameter corresponding to the Large dot. In this case, in S708 and S716 of Small quantization processing S662, the image processing device 12 sets the output value to a value corresponding to the Small dot.

  In this case, in the middle quantization process S652 and the small quantization process S662 performed after the large quantization process S642, the output value is set to a value other than 0 in the previously performed process. The determination result in S702 changes depending on whether or not there is. As a result, in the middle quantization process S652 and the small quantization process S662, the image processing apparatus 12 changes the process to be performed depending on whether or not the output value is set to a value other than 0. . Hereinafter, this point will be described in more detail.

  For example, the processing result of Large quantization processing S642 is a result corresponding to the case of forming a dot (the dot is in an ON state), and the image processing apparatus 12 sets the output value to a value corresponding to the Large dot. In this case, in step S702 in the middle quantization process S652 and the small quantization process S662, the image processing apparatus 12 determines that the output value is other than 0, and proceeds to step S704. Omitted. In this case, the processing results of the Middle quantization processing S652 and the Small quantization processing S662 are forced to be the results corresponding to the case where dots are not formed (the dot is in an OFF state).

  Further, if the processing result of the Large quantization process S642 is a result (OFF) corresponding to the case where no dot is formed, and the output value remains 0 even after the Large quantization process S642, the result for Middle In the quantization process S652, the image processing apparatus 12 proceeds to step S706 and subsequent steps, and performs the same process as the large quantization process S642 using the error-corrected input value for middle dots and the noise-corrected threshold value. As a result, the image processing device 12 determines ON / OFF of the middle dot.

  When the Middle dot quantization result is turned ON, the Small dot quantization result is forcibly turned OFF in the same manner as described above. When the quantization result of the Middle dot is OFF, in the Small quantization process S662, the image processing apparatus 12 proceeds to Step S706 and subsequent steps, and sets the error corrected input value and the noise corrected threshold value for the Small dot. The same process as the Large quantization process S642 is performed. Thereby, the image processing apparatus 12 determines ON / OFF of the Small dot.

  Therefore, in this example, for example, in each stage determination process (Large quantization process S642, Middle quantization process S652 or Small quantization process S662) corresponding to any dot size, this process is supported. When it is determined that a dot of an ink droplet having a dot size to be formed is formed, in the step-by-step determination process corresponding to a dot size smaller than the dot size, the image processing device 12 forms a dot regardless of the density value of the pixel. Judge that not. For example, when it is determined that a dot is to be formed in any of the determination processes for each stage, the image processing apparatus 12 performs quantization corresponding to the maximum dot size among the determination processes for each stage determined to form dots. The result value is determined as the quantization result.

  If it is not determined that dots are formed in any of the determination processes for each stage, the image processing apparatus 12 determines a value corresponding to the case where dots are not formed as a result of quantization. Accordingly, in the quantization execution process S612, the image processing apparatus 12 performs discrete processing corresponding to each dot size based on the results of the Large quantization process S642, the Middle quantization process S652, and the Small quantization process S662. One of the typical values is determined as the quantization result and set as the output value. According to this example, multi-tone quantization processing can be performed appropriately.

  For example, when multi-tone quantization is performed by a conventional method, dots of ink droplets having different dot sizes may be mixedly formed depending on the pixel density. For example, as in this example, when using three types of dot sizes, Large dots, Middle dots, and Small dots, printing is performed using only Small dots in a range where the pixel density is small. When the density is increased, Small dots and Middle dots are mixedly formed. In this case, at a density at which dots of different sizes begin to mix, for example, when viewed only with a larger size of middle dots, it corresponds to the density value of the highlight portion. In this case, in the quantization process, for example, if only the error diffusion process is performed, a dot delay occurs and the image quality deteriorates. For this reason, in multi-tone quantization, it is desirable to appropriately eliminate the dot delay that occurs at the part where dots of different sizes begin to mix.

  On the other hand, in this example, for example, in each of the Large quantization process S642, the Middle quantization process S652, and the Small quantization process S662, the output value is 0 at the start of processing. The image processing apparatus 12 performs a quantization process between two gradations of 0, which is the lowest quantization value, and a value corresponding to the corresponding dot size. Also, in this two-tone quantization process, dither matrix noise is used in accordance with the noise usage rate corresponding to the pixel to be processed, and the accumulated error is in accordance with the error usage rate corresponding to the pixel. Is used. Therefore, according to this example, for example, even when performing multi-gradation quantization, for example, a quantization process that makes good use of a portion that each process of the dither process and the error diffusion process is good at is appropriately performed. It can be carried out.

  In this case, since the occurrence of dot delay in the highlight portion can be appropriately suppressed, for example, the occurrence of dot delay can be appropriately eliminated even in a portion where dots of different sizes start to mix. Thereby, for example, the change in dot size can be smoothed. Furthermore, unlike the case where the dither processing and the error diffusion processing are simply switched according to the density value, no boundary line is generated at the switching portion between the dither processing and the error diffusion processing. Therefore, according to this example, for example, multi-tone quantization can be performed more appropriately. Thereby, for example, the dot size to be formed can be appropriately set according to the density value of the pixel. In addition, multi-tone quantization can be performed more appropriately, and high-definition, high-quality images can be printed appropriately.

  FIG. 16 is a flowchart showing a modification of the process in the quantization execution process S612. Except for the points described below, in FIG. 16, steps denoted by the same reference numerals as those in FIG. 14 perform the same or similar processing as the steps in FIG.

  In this example, the image processing apparatus 12 sets an output value indicating the result of multi-gradation quantization separately from the Large quantization process S642, the Middle quantization process S652, and the Small quantization process S662. In the steps of

  For example, in the large process S640, the image processing apparatus 12 performs the large quantization process S642 and the large error distribution process S644, and then determines the result of the large quantization process S642 (quantization result determination). Process S646). If the quantization result for the Large dot is OFF, that is, the error-corrected input value is equal to or less than the noise-corrected threshold value (S646: Yes), the next Middle processing is performed without changing the output value. Proceed to S650. When the quantization result for the Large dot is ON, that is, when the error-corrected input value is larger than the noise-corrected threshold (S646: No), the output value is set to a value corresponding to the Large dot (output value). The setting process S648) proceeds to the next Middle process S650.

  In the large process S640 in this example, the image processing apparatus 12 has determined whether the quantization result is ON or OFF instead of changing the output value as in the process described with reference to FIG. Is transferred to the subsequent processing. In the quantization result determination process S646, the image processing apparatus 12 determines the quantization result based on this information.

  In the middle processing S650, the image processing apparatus 12 performs the middle quantization processing S652, the middle error distribution processing S654, the quantization result determination processing S656, and the output value setting processing S658 in the same manner as the large processing S640. . The middle quantization process S652, the middle error distribution process S654, the quantization result determination process S656, and the output value M setting process S658 each replaces each parameter corresponding to the middle dot with each parameter corresponding to the large dot. Except for the use, the processing is the same as or similar to the Large quantization processing S642, the Large error distribution processing S644, the quantization result determination processing S646, and the output value setting processing S648.

  In the Small process S660, the image processing apparatus 12 performs the Small quantization process S662, the Small error distribution process S664, the quantization result determination process S666, the output value setting process S668, in the same manner as the Large process S640. And the output value 0 setting process S670 is performed. Among these, each of the Small quantization process S662, the Small error distribution process S664, the quantization result determination process S666, and the output value setting process S668, each corresponding to the Small dot instead of each parameter corresponding to the Large dot. Except for the use of parameters, the processing is the same as or similar to the Large quantization processing S642, the Large error distribution processing S644, the quantization result determination processing S646, and the output value setting processing S648.

  In the quantization result determination process S666 of the Small process S660, when the quantization result for the Small dot is OFF (S666: Yes), the image processing apparatus 12 proceeds to the output value 0 setting process S670, and outputs it. The value is set to 0, which is a value corresponding to the case where no dot is formed. Also in this example, multi-tone quantization processing can be performed appropriately.

  FIG. 17 is a flowchart showing a second example of the multi-tone quantization process, and shows a modification of the process executed between steps S604 and S614 in FIG. Except for the points described below, in FIG. 17, steps denoted by the same reference numerals as in FIGS. 11 to 16 perform the same or similar processing as the steps in FIGS. 11 to 16.

  In this example, the image processing apparatus 12 performs a pixel selection process S606 and an input value calculation process S608 following S604. These processes may be, for example, the same processes as the pixel selection process S606 and the input value calculation process S608 described with reference to FIG. Further, for example, in the input value calculation process S608, the image processing apparatus 12 may calculate the L curve value, the M curve value, the S curve value, and the like using a dot size determination table different from the process of FIG.

  Subsequently, the image processing apparatus 12 performs a large process S640, a middle process S650, and a small process S660 in parallel. These processes are quantization and error distribution processes corresponding to each dot size. In these processes, the image processing apparatus 12 performs a quantization process corresponding to each dot size, similarly to the large process S640, the middle process S650, and the small process S660 described with reference to FIG. Perform error distribution processing.

  In this quantization process, for example, the image processing apparatus 12 performs the same process as the Large quantization process S642 described with reference to FIG. However, in this case, for example, the image processing apparatus 12 omits steps S702 and S704 in FIG. 15 and starts the processing from step S706. In the processing corresponding to S708 and S716, for example, instead of directly changing the output value, the image processing apparatus 12 sets a value corresponding to each dot size in another parameter indicating the output result of the quantization processing. Set.

  In this example, the value corresponding to each dot size is a value corresponding to each of a large dot, a middle dot, and a small dot. In this example, the value corresponding to the Large dot is larger than the value corresponding to the Middle dot. The value corresponding to the Middle dot is a value larger than the value corresponding to the Small dot. The value corresponding to the Small dot is a value larger than 0 which is a value indicating a state where no dot is formed. In the large process S640, the middle process S650, and the small process S660, the image processing apparatus 12 performs, for example, the large error distribution process S644, the middle error distribution process S654, and the small error distribution process in FIG. Similar to S664, error distribution processing is performed.

  After these parallel processes, the image processing apparatus 12 proceeds to S682 for performing the next process, and the output value initialized to 0 in the input value calculation process S608, the output result of the quantization process in the large process S640, and Are compared (S682). If the output result of the quantization process is larger than the output value (S682: Yes), the output value is set to a value corresponding to the Large dot (S684), and the process proceeds to the next step S614.

  In S682, when the output result of the quantization process is equal to or less than the output value (S682: No), the image processing apparatus 12 compares the output value with the output result of the quantization process in the Middle process S650. (S686). If the output result of the quantization process is larger than the output value (S686: Yes), the output value is set to a value corresponding to the Middle dot (S688), and the process proceeds to the next step S614.

  In S686, when the output result of the quantization process is equal to or less than the output value (S686: No), the image processing apparatus 12 compares the output value with the output result of the quantization process in the Small process S650. (S690). If the output result of the quantization process is larger than the output value (S690: Yes), the output value is set to a value corresponding to the Small dot (S692), and the process proceeds to the next step S614. In S690, if the output result of the quantization process is equal to or less than the output value (S690: No), the image processing apparatus 12 proceeds to the next step S614 while keeping the output value at 0, which is the initialization value. move on.

  When configured in this manner, the image processing apparatus 12 performs the two-tone quantization process as a part of the multi-tone quantization process as a large dot process S640, a middle dot process S650, and This is executed in each of the small dot processing S660. Then, after the processing for each dot size, the comparison based on the output result of each processing is sequentially performed, and the value corresponding to the largest dot size among the dot sizes in which the dots are in the ON state is calculated. Set to the output value of the key quantization process. Further, when the result corresponding to all dot sizes results in no dots being placed (OFF), the value 0 corresponding to the state in which no dots are formed is set as the output value. In this example as well, for example, multi-tone quantization processing can be appropriately performed by making use of the advantages of the dither processing and error diffusion processing.

  Hereinafter, further modifications of the present invention will be described. FIG. 18 is a flowchart illustrating an example of a process for correcting the calculated noise usage rate. The image processing apparatus 12 executes the process of this flowchart before the quantization process in, for example, the Large dot process S640, the Middle dot process S650, and the Small dot process S660.

  Note that the process described below can be applied to the multi-tone quantization process described with reference to FIG. Further, for example, the present invention may be applied to the multi-gradation quantization process described with reference to FIGS.

  When two-tone quantization processing (hybrid error diffusion processing) using both dither matrix noise and accumulated error is performed for each dot size, for example, as described with reference to FIG. Depending on the error, the error usage rate and the noise usage rate are different, and the dot arrangement may be slightly different. And as a result, interference arises and the part where a dot becomes a ball shape in a printing result may arise.

  In response to this, for example, by performing a correction that reduces the difference in noise usage rate for each dot size, the occurrence of a ball-shaped portion can be suppressed. This correction process is executed by, for example, varying the noise usage rate according to the curve value corresponding to the dot size other than the dot size being processed with respect to the noise usage rate calculated from a preset relational expression. it can.

  More specifically, for example, when the noise usage rate is corrected in the large dot processing S640, the image processing apparatus 12 first performs the noise described with reference to FIG. 13 as shown in the flowchart of FIG. It is determined whether the noise usage rate is the maximum value with respect to the noise usage rate calculated in the usage rate calculation processing S620 (S802). If it is not the maximum value (S802: No), for each of the M curve value and S curve value, which are input values for each stage corresponding to each of the Middle dot and Small dot, from the shadow-side noise use rate maximum density value Sn. Is also larger (S804).

  If any of the curve values is larger than the shadow-side noise usage rate maximum density value Sn (S804: Yes), a preset usage rate addition value RnAdd is added to the noise usage rate corresponding to the Large dot. . As a result, when the curve value other than the dot size being processed is larger than the shadow-side noise usage rate maximum density value Sn, the image processing apparatus 12 sets the noise usage rate calculated in the noise usage rate calculation processing S620 to a slightly larger value. (S806). Then, the corrected noise usage rate is determined as a new noise usage rate used in the quantization process (S808).

  If it is determined in S802 that the noise usage rate is the maximum value (S802: Yes), the image processing apparatus 12 does not correct the noise usage rate and converts the noise usage rate as it is by the quantization process. The noise usage rate to be used is determined (S808). When correcting the noise usage rate in the middle dot processing S650, in S804, the image processing apparatus 12 determines that either the L curve value or the S curve value is greater than the shadow side noise usage rate maximum density value Sn. Determine whether it is larger. When the noise usage rate is corrected in the small dot processing S660, in S804, the image processing apparatus 12 determines whether either the L curve value or the M curve value is larger than the shadow side noise usage rate maximum density value Sn. Determine whether or not.

  In this case, for example, when the error usage rate is large and the noise usage rate is low for the dot size being processed, and the noise usage rate is high for other dot sizes, the noise of the dot size being processed The usage rate can be appropriately corrected to a slightly large value. Also, by using a slightly larger value of the noise usage rate, the difference in noise usage rate corresponding to each dot size is reduced and uniformized. Therefore, according to this example, it is possible to appropriately suppress, for example, a difference in dot arrangement for each dot size. Moreover, it can suppress appropriately that the part from which a dot becomes a ball shape arises in a printing result by this.

  FIG. 19 shows an example of a range for correcting the noise usage rate together with an example of a dot size determination table. The dot size determination table described below is a modification of the dot size determination table described with reference to FIG. 12, and can be used, for example, in the multi-tone quantization process described with reference to FIG. Further, these dot size determination tables may be used in the multi-tone quantization process described with reference to FIGS.

  FIG. 19A shows an example of a dot size determination table. FIG. 19B shows another example of the dot size determination table. The dot size determination table can be variously changed according to, for example, the performance of the printing apparatus, the contents of the image forming process, or the desired printing result. Therefore, in the multi-tone quantization process described with reference to FIG. 17, a dot size determination table as shown in FIG. 19A is used instead of the dot size determination table shown in FIG. You can also

  In this case, for example, in the input value calculation process S608 in FIG. 17, the image processing apparatus 12 sets the value indicated by the large dot curve to the L curve value according to the density value that is the input value. Further, a value obtained by adding the value indicated by the Middle dot curve and the L curve value is set as the M curve value. A value obtained by adding the value indicated by the Small dot curve and the M curve value is set as the S curve value.

  Further, as the dot size determination table, for example, a dot size determination table as shown in FIG. 19B can be used. In this case, for example, in the input value calculation process S608 in FIG. 17, the image processing apparatus 12 calculates the L curve value, the M curve value, and the S curve value in the same or similar manner as the input value calculation process S608 in FIG. To do.

  Also in the case described above, the L curve value, the M curve value, and the S curve value can be calculated appropriately. In these cases, the image processing apparatus 12 corrects the noise usage rate in a range surrounded by a shaded square in each figure. In this way, for example, it is possible to appropriately correct the noise usage rate. This also makes it possible to perform high-quality printing more appropriately.

  As mentioned above, although this invention was demonstrated using embodiment, the technical scope of this invention is not limited to the range as described in the said embodiment. It will be apparent to those skilled in the art that various modifications or improvements can be added to the above embodiment. It is apparent from the description of the scope of claims that embodiments with such changes or improvements can be included in the technical scope of the present invention.

  The present invention can be suitably used for a program that causes a computer to perform quantization, for example.

DESCRIPTION OF SYMBOLS 10 ... Printing system, 12 ... Image processing apparatus, 14 ... Printing apparatus

Claims (8)

A program for causing a computer to perform a multi-tone quantization process, which is a process of converting a density value indicating the color density of each pixel in an image into a discrete value of three or more stages,
A pixel selection process for sequentially selecting the pixels to be quantized;
Accumulate a noise usage rate indicating the degree of influence of the dither matrix noise, which is noise specified by a preset dither matrix, on the quantization process, and a quantization error that is an error generated in the quantization of surrounding pixels. And determining an error usage rate indicating a degree of influence of the accumulated error on the quantization process, and according to the density value of pixels sequentially selected in the pixel selection process, the pixel corresponding to the pixel A usage rate determination process for determining a noise usage rate and the error usage rate;
A process for performing the quantization on the density values of the pixels sequentially selected in the pixel selection process, using dither matrix noise according to the noise usage rate corresponding to the pixel, and corresponding to the pixel And causing the computer to execute a quantization execution process for performing the quantization using the accumulated error according to the error usage rate,
The usage rate determination process sets the noise usage rate to a value greater than or equal to a minimum noise usage rate preset to a value greater than 0, regardless of the density value of the pixel .
The noise usage rate is gradually reduced with respect to the change in the density value of the pixel in the direction from the highlight portion toward the halftone portion between the highlight portion and the halftone portion, and the error is increased. Gradually increase the usage rate,
Between the halftone portion and the shadow portion, the noise usage rate is gradually increased with respect to a change in the density value of the pixel in the direction from the halftone portion to the shadow portion, and the error use is performed. A program characterized by gradually decreasing the rate . The discrete values of the three or more levels are quantized by a minimum quantized value that is a value corresponding to a result of quantizing the minimum density value and a value that is larger than the minimum density value by a predetermined amount or more. A value corresponding to the result, and including at least two or more stages of quantization result values that are discrete values larger than the lowest quantization value,
The quantization execution process is a process performed corresponding to each of the two or more stages of quantization result values, and whether or not the result of quantizing the density value should be a value greater than or equal to the quantization result value. Including stage-by-stage judgment processing,
The determination processing for each step performed corresponding to at least one of the two or more quantization result values uses the dither matrix noise according to the noise usage rate corresponding to the pixel to be processed. And, using the accumulated error according to the error usage rate corresponding to the pixel, quantization of two gradations between the two gradations of the lowest quantization value and the quantization result value Process
In the quantization execution process, one of the discrete values of the three or more stages is selected as the pixel selection based on the result of the determination process for each stage corresponding to each of the quantization result values of the two or more stages. 2. The program according to claim 1, wherein the program is determined as a result of the multi-gradation quantization with respect to a density value of pixels sequentially selected in processing. The program causes the computer to perform the multi-gradation quantization during an image forming process for forming printable data indicating an image in a format that can be interpreted by a printing apparatus,
The printing apparatus is an inkjet printer that performs multi-tone printing by ejecting a plurality of types of ink droplets having different dot sizes.
Each of the two or more stages of quantization result values is associated with the respective dot sizes of the plurality of types of ink droplets,
The step-by-step determination process determines whether or not to form dots of ink droplets of the dot size corresponding to the step-by-step determination process by performing the two-tone quantization process.
If it is not determined to form the dot in any of the determination processes for each stage, the quantization execution process determines the minimum quantization value as a result of the multi-gradation quantization,
When it is determined in any one of the determination processes for each step that the dots are formed, the quantization execution process corresponds to the maximum dot size among the determination processes for each step determined to form the dots. The program according to claim 2, wherein the quantization result value is determined as a result of the multi-tone quantization. The quantization execution process first performs the determination process for each stage corresponding to a larger dot size,
In the step-by-step determination process corresponding to any one of the dot sizes, when it is determined to form a dot of an ink droplet having the dot size, the step-by-step determination process corresponding to the dot size smaller than the dot size is performed. The program according to claim 3, wherein it is determined that the dot is not formed regardless of the density value of the pixel. The program is a process of calculating an input value for each stage, which is a value used as an input value in the determination process for each stage corresponding to each of the dot sizes, and the density of the pixels sequentially selected in the pixel selection process In accordance with the value, based on preset input value calculation data, the computer further executes an input value calculation process for calculating the input value for each stage corresponding to each dot size,
The input value calculation data for each stage is data that determines the number of dots of ink droplets to be formed per unit area according to the density value of the pixel for each dot size.
In the input value calculation process, the input value for each step corresponding to each dot size is set in the input value calculation data for the density value of the pixels sequentially selected in the pixel selection process. Is calculated according to the number of dots associated with the density value,
When performing the two-gradation quantization, the step-by-step determination processing corresponding to each dot size performs the two-gradation quantization processing on the step-by-step input value corresponding to the dot size. The program according to claim 3 or 4, wherein the program is executed. The quantization execution process further includes a stage-by-stage cumulative error calculation process for calculating the cumulative error corresponding to the quantization result value corresponding to each of the two or more stages of quantization result values.
The step-by-step determination process uses the cumulative error corresponding to the quantization result value that is the same as the step-by-step determination process as the cumulative error,
6. The cumulative error calculation process for each stage updates the cumulative error corresponding to the quantization result value based on a result of the process by the determination process for each stage. Program. An image processing apparatus that performs multi-tone quantization processing, which is processing for converting density values indicating the color density of each pixel in an image into discrete values of three or more stages,
A pixel selection processing unit that sequentially selects the pixels to be quantized;
Accumulate a noise usage rate indicating the degree of influence of the dither matrix noise, which is noise specified by a preset dither matrix, on the quantization process, and a quantization error that is an error generated in the quantization of surrounding pixels. A processing unit for determining an error usage rate indicating a degree of affecting the quantization process, and corresponding to the pixel according to the density value of pixels sequentially selected by the pixel selection processing unit. A usage rate determination processing unit for determining the noise usage rate and the error usage rate;
The processing unit that performs the quantization on the density values of the pixels sequentially selected by the pixel selection processing unit, uses dither matrix noise according to the noise usage rate corresponding to the pixel, and the pixel A quantization execution processing unit that performs the quantization using the accumulated error according to the error usage rate corresponding to
The usage rate determination processing unit sets the noise usage rate to a value equal to or higher than a minimum noise usage rate preset to a value larger than 0 regardless of the density value of the pixel .
The noise usage rate is gradually reduced with respect to the change in the density value of the pixel in the direction from the highlight portion toward the halftone portion between the highlight portion and the halftone portion, and the error is increased. Gradually increase the usage rate,
Between the halftone portion and the shadow portion, the noise usage rate is gradually increased with respect to a change in the density value of the pixel in the direction from the halftone portion to the shadow portion, and the error use is performed. An image processing apparatus characterized by gradually reducing the rate . An image processing method for performing a multi-tone quantization process, which is a process of converting a density value indicating a color density of each pixel in an image into a discrete value of three or more stages,
A pixel selection processing step of sequentially selecting the pixels to be quantized;
Accumulate a noise usage rate indicating the degree of influence of the dither matrix noise, which is noise specified by a preset dither matrix, on the quantization process, and a quantization error that is an error generated in the quantization of surrounding pixels. And an error usage rate indicating a degree of influence of the accumulated error on the quantization process, and corresponding to the pixel according to the density value of the pixels sequentially selected in the pixel selection process stage. A usage rate determination processing step for determining the noise usage rate and the error usage rate;
A step of performing the quantization on the density values of the pixels sequentially selected in the pixel selection processing step, using dither matrix noise according to the noise usage rate corresponding to the pixel, and the pixel A quantization execution processing stage that performs the quantization using the accumulated error in accordance with the error usage rate corresponding to:
The usage rate determination processing step sets the noise usage rate to a value equal to or higher than a minimum noise usage rate preset to a value larger than 0 regardless of the density value of the pixel .
The noise usage rate is gradually reduced with respect to the change in the density value of the pixel in the direction from the highlight portion toward the halftone portion between the highlight portion and the halftone portion, and the error is increased. Gradually increase the usage rate,
Between the halftone portion and the shadow portion, the noise usage rate is gradually increased with respect to a change in the density value of the pixel in the direction from the halftone portion to the shadow portion, and the error use is performed. An image processing method characterized by gradually reducing the rate .

Images