JP4577366B2 - Image processing apparatus for processing a plurality of rasters in parallel - Google Patents

Description

  The present invention relates to a technique for converting image data in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally, and more specifically, a pixel array while judging the presence or absence of dot formation for each pixel. The present invention relates to a technique for converting image data into image data in an expression format depending on whether or not dots are formed by converting a raster into a dot row.

  An image display device that expresses an image by forming dots on a display medium such as a print medium or a liquid crystal screen is widely used as an output device of various image devices. Such an image display device can express only the state of whether or not to form dots locally, but by appropriately controlling the dot formation density according to the gradation value of the image, It is possible to express an image whose tone changes continuously.

  In these image display devices, various methods have been proposed as methods for determining the presence or absence of dot formation for each pixel so that dots are formed at an appropriate density according to the gradation value of the image. As a typical method, there are a method called a dither method and a method called an error diffusion method.

  The error diffusion method is a method of calculating the error of gradation expression that occurs when dots are formed or not formed on a pixel, and judging the presence or absence of dot formation while taking into account the error generated in surrounding pixels. It is. If the error diffusion method is employed, it is possible to display image data with high accuracy by determining the presence or absence of dot formation so as to eliminate errors generated in surrounding pixels. On the other hand, the dither method does not require processing for calculating an error in gradation expression, and can determine whether or not dots are formed without considering the error generated in the surrounding pixels. High-speed processing is possible. From this, it is possible to display quickly by applying the dither method when priority is given to display speed based on the balance between image display speed and display image quality. By applying the diffusion method, an image with good image quality can be displayed.

  However, in recent years, the number of pixels constituting image data tends to increase gradually, and the time required for image data conversion also increases accordingly. Therefore, even when the dither method is applied, the image is displayed quickly. There are cases where this is difficult. Therefore, in order to enable quick display even for image data having a large number of pixels, development of a more rapid image processing method is required.

  The present invention has been made in order to solve the above-described problems in the prior art. By reducing the time required for determining whether or not dots are formed without deteriorating the image quality, a high-quality image can be quickly obtained. The purpose is to provide technology that can be displayed on the screen.

In order to solve at least a part of the above-described problems, the first image processing apparatus of the present invention employs the following configuration. That is,
Receiving image data in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally, comparing the gradation value of each pixel with a predetermined threshold value and determining whether or not a dot is formed An image processing apparatus that converts the image data into data in an expression format depending on whether or not dots are formed by converting a raster as a row of dots into a dot row,
Threshold matrix storage means for storing a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged;
Target pixel selection means for selecting a target pixel for determining whether or not to form a dot;
A dot formation determination unit that reads out a threshold value corresponding to the target pixel from the threshold value matrix and compares the threshold value with the gradation value corresponding to the target pixel, thereby determining whether or not the target pixel is formed. And
The target pixel selection means collects a predetermined N number of adjacent rasters (N is an integer of 2 or more) as a raster group, and sequentially selects each pixel included in the raster group according to a predetermined order, The gist is that the target pixel is selected from the N rasters in parallel.

Further, the first image processing method of the present invention corresponding to the first image processing apparatus is as follows.
Receiving image data in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally, comparing the gradation value of each pixel with a predetermined threshold value and determining whether or not a dot is formed An image processing method for converting the image data into data in an expression format depending on whether or not dots are formed by converting a raster as a row of dots into a dot row,
A first step of preliminarily storing a threshold value matrix in which a plurality of the threshold values are two-dimensionally arranged;
A second step of selecting a target pixel for determining whether or not to form a dot;
A third step of determining the presence or absence of dot formation for the target pixel by reading out the threshold corresponding to the target pixel from the threshold matrix and comparing it with the gradation value associated with the target pixel And
In the second step, predetermined N (N is an integer of 2 or more) adjacent rasters are grouped as a raster group, and each pixel included in the raster group is sequentially selected according to a predetermined order, The gist is that the target pixel is selected from the N rasters in parallel.

  In the first image processing apparatus and image processing method, the raster is converted into a dot row by determining the presence or absence of dot formation for sequentially selected target pixels. When selecting the target pixel, the predetermined N adjacent (N is an integer of 2 or more) rasters are grouped into a raster group, and each pixel included in the raster group is sequentially selected according to a predetermined order. The target pixel is selected from each raster while N rasters are arranged in parallel. As a result, the image data is converted into data in an expression format depending on whether or not dots are formed, while converting the rasters N by dot into parallel. If N rasters are converted into dot rows in parallel in this way, the conversion can be performed quickly for the following reason.

  In general, image data tends to be associated with approximate gradation values between neighboring pixels. Further, the same gradation value is often associated with a plurality of adjacent pixels. Therefore, when converting such image data using an image processing apparatus, it is possible to convert the image data quickly by using a cache memory of a CPU built in the image processing apparatus. Here, the cache memory is a high-speed readable / writable memory provided in a storage hierarchy immediately below the CPU. The CPU advances the process while generating the execution code one after another inside, and temporarily stores the generated execution code in the cache memory. If the gradation values of pixels to be continuously processed are approximate, there are many cases where the execution code used in the processing of the previous pixel and remaining in the cache memory can be used as it is. If the execution code in the cache memory is used, there is no need to newly generate an execution code, so that it is possible to perform the processing as quickly as that.

  Here, since the pixels are two-dimensionally arranged, it is considered that the gradation values that are similarly approximated are associated with pixels that are adjacent in any direction. That is, when attention is paid to a certain pixel, it is considered that the approximate gradation value is associated not only with the pixel adjacent to the pixel in the horizontal direction but also with the pixel adjacent in the vertical direction or diagonal direction. Therefore, if a plurality of rasters are processed in parallel, rather than processing rasters one by one, continuous processing is performed not only in the left-right direction but also in pixels with approximate gradation values that exist in the vertical or diagonal direction. can do. As a result, the execution code remaining in the cache memory can be used efficiently, so that it is possible to perform the processing more quickly.

  In recent years, various CPUs that internally perform pipeline processing have been developed and used. Various CPUs that perform branch prediction have been developed and used in order to perform pipeline processing more efficiently. Pipeline processing is the following processing. When the CPU executes individual instructions described in the program, the CPU executes it in several steps as follows. First, at least one instruction is read, such as (1) reading an instruction, (2) interpreting the read instruction, (3) executing the interpreted content, and (4) returning the obtained result. The process is divided into four steps. Pipeline processing is processing in which a dedicated unit corresponding to these procedures is provided in the CPU, so that one instruction flows and is executed in a work manner. For example, the next instruction is read while the read instruction is being interpreted, and the next instruction is interpreted while the interpreted instruction is being executed. Of course, the next command may be read while the next command is being interpreted. In this way, if the processing is carried out in a flow manner, the command can be executed very quickly.

  In the pipeline processing, when a branch instruction is executed, it is useless to interpret the next instruction or read the next instruction, so that the processing speed decreases. However, if the gradation values of pixels to be continuously processed are approximated, the same processing is repeatedly executed for those pixels, and therefore, the possibility that a branch for switching the processing will occur is reduced. For this reason, it is possible to perform image processing quickly.

  Furthermore, even if a branch instruction is generated during the processing of a certain pixel, it is considered that a branch instruction is generated during the processing of subsequent pixels as long as the gradation values of pixels to be continuously processed are approximate. The branch instruction is generated according to a certain pattern. In such a case, it is possible to predict the occurrence of a branch instruction at a high hit rate, so that the work of reading and interpreting the subsequent instruction is avoided and higher speed processing is possible. It becomes.

  For this reason, if a plurality of rasters are processed in parallel, pixels that are two-dimensionally adjacent and approximate to gradation values can be processed in succession. In addition, the efficiency of pipeline processing can be improved, and the accuracy of branch prediction can be further improved. As a result, it is possible to quickly convert the image data into data expressed by the presence or absence of dot formation.

  In such an image processing apparatus, the plurality of threshold values may be stored in a matrix of M rows (M is a positive multiple of N) × L columns (L is an integer of 2 or more).

  If the threshold value matrix is set to such a size, when the image data is divided into raster groups each including the N rasters, rasters included in one raster group can always be associated with the same threshold matrix. . In other words, when the threshold matrix is associated with the image data, there is no threshold matrix boundary between the rasters constituting the raster group. For this reason, it is possible to increase the speed of image processing while avoiding complicated processing for reading out the threshold value corresponding to the target pixel from the threshold value matrix.

  In these image processing apparatuses or image processing methods described above, the threshold values of the threshold matrix may be stored as follows. First, a plurality of threshold values arranged in a matrix are grouped into a plurality of groups by the N rows. In each group, the target pixel selection unit stores the plurality of threshold values in the predetermined order in which the target pixel is selected.

  In this way, each threshold value of the threshold value matrix can be stored in accordance with the order of selecting the target pixel from the N rasters constituting the raster group. As a result, when the gradation value of the image data for the target pixel and the threshold value corresponding to the target pixel are read to determine the presence / absence of dot formation, the threshold values may be read according to the stored order. Accordingly, the processing becomes simple and the processing speed can be improved. Note that “the target pixel selection means stores a plurality of thresholds according to a predetermined order in which the target pixels are selected” means that the order in which the target pixels are selected and the order in which the plurality of thresholds are stored are completely Not only those that match, but also those that are somewhat out of order. That is, as long as the order in which the target pixel is selected and the order in which the threshold value is stored are slightly different from each other, the threshold value corresponding to the target pixel can be easily found from the threshold value matrix. It is because it can be made simple. Further, if each threshold value of the threshold value matrix is stored in accordance with the order of selecting the target pixels in this way, for example, when the CPU that performs image processing does not have a built-in cache or when pipeline processing is not performed. Even if it exists, it becomes possible to improve a processing speed.

  In such an image processing apparatus, the target pixel may be selected as follows. A plurality of pixels included in the raster group are grouped into a plurality of blocks each having a predetermined number of pixels in the vertical and horizontal directions, and the target pixel is sequentially selected from each pixel in the block.

  By doing this, it is possible to sequentially select the target pixels from the pixels constituting the raster group by repeatedly performing the process of selecting the target pixels from each pixel in the block. As a result, the processing can be simplified, and the image processing speed can be improved.

  In addition, such a block may be a block in which a plurality of pixels each having N pixels in length and width are combined. In this case, the target pixel can be sequentially selected from the pixels constituting the raster group simply by moving the block sequentially in the raster direction, so that the processing can be further simplified.

  Such a block can be a block in which a total of 4 pixels, particularly 2 pixels vertically and horizontally, are combined. Since each block of 2 pixels in the vertical and horizontal directions has a small number of pixels, the processing is not complicated, and the image processing speed can be effectively improved, which is preferable.

  Alternatively, a total of 64 pixels of 8 pixels in the vertical and horizontal directions can be combined into a block. For example, when image data compressed in the JPEG format is decompressed, 8 pixels × 8 pixels are grouped and developed into gradation value data of each pixel. Therefore, by configuring a block of 8 pixels in length and width and selecting a target pixel from such a block, image processing can be performed in the same order without rearranging the developed gradation values. Processing can be speeded up.

  In the above-described image processing apparatus or image processing method, the presence / absence of dot formation for the target pixel may be determined as follows. First, an array having a size smaller than the two-dimensional pixel array of the image data is stored as the threshold value matrix. Then, the threshold value matrix is arranged by shifting K pixels (K is a positive multiple of N) in a direction intersecting the raster direction from a state in which the threshold value matrix is arranged in a vertical and horizontal grid shape with respect to the image data. And the threshold value stored in the position corresponding to the target pixel is read out from the threshold value matrix associated with the image data. The presence / absence of dot formation may be determined by comparing the threshold value thus read out with the gradation value of the target pixel.

  In this way, if the threshold value matrix is associated with the image data while being shifted little by little, it is possible to avoid the same threshold value being repeatedly used in the raster direction and repeating the formation of dots in a certain pattern. Further, if the shift amount of the threshold matrix is set to K pixels (K is a positive multiple of N) in a direction crossing the raster direction, a raster group is formed when the threshold matrix is associated with image data. There is no threshold matrix boundary between rasters. For this reason, it is possible to increase the speed of image processing while avoiding complicated processing for reading out the threshold value corresponding to the target pixel from the threshold value matrix.

  When the threshold value matrix is thus shifted and associated with the image data, the shift amount may be set to the same value as the number of rasters constituting the raster group. As will be described in detail later, this is preferable because the process of reading out each threshold value of the threshold value matrix can be simplified.

In order to solve at least a part of the problems described above, the second image processing apparatus of the present invention employs the following configuration. That is,
Receiving color image data in which gradation values of a plurality of predetermined colors are associated with each of a plurality of pixels arranged two-dimensionally, and determining the presence or absence of dot formation for each pixel for each color component, An image processing apparatus that converts color image data into data of each color in an expression format depending on whether or not dots are formed by converting a raster line to a dot line,
Raster group generating means for generating a raster group by combining predetermined N (N is an integer of 2 or more) adjacent rasters;
For the image data of the first color component selected in advance from the color components, the raster group is configured by comparing the gradation value of each pixel with a predetermined threshold value associated with each pixel. First data conversion means for converting the N rasters to dot rows in parallel, thereby converting the data into an expression format depending on whether or not dots are formed;
For the image data of the second color component obtained by removing the first color component from each color component, the gradation error caused by the pixel determined by determining the presence / absence of dot formation is diffused to surrounding undetermined pixels. At the same time, the N rasters constituting the raster group are converted into dot rows in parallel while judging the presence / absence of dot formation so as to eliminate the gradation error diffused from the surrounding judged pixels. And a second data conversion means for converting the data into an expression format data based on the presence or absence of dot formation.

The second image processing method of the present invention corresponding to the second image processing apparatus is as follows.
Receiving color image data in which gradation values of a plurality of predetermined colors are associated with each of a plurality of pixels arranged two-dimensionally, and determining the presence or absence of dot formation for each pixel for each color component, An image processing method for converting color image data into data of each color in an expression format depending on the presence or absence of dot formation by converting a raster that is a row into a dot row,
Separating the received color image data into image data of a first color component preselected from the color components and image data of a second color component excluding the first color component; When,
For the image data of the first color component, a predetermined number of adjacent rasters (N is 2 or more) are compared while comparing the gradation value of each pixel with a predetermined threshold value associated with each pixel. A step of converting the data into an expression format according to the presence or absence of dot formation,
As for the image data of the second color component, the gradation error generated in the pixel determined by determining the presence / absence of dot formation is diffused to the surrounding undetermined pixels and diffused from the surrounding determined pixels. By converting the adjacent rasters into dot rows in parallel while determining the presence / absence of dot formation so as to eliminate the tone error, the data is converted into data in an expression format according to the presence / absence of dot formation. The gist is to provide a process.

  In the second image processing apparatus and the image processing method, the gradation value of each pixel and each pixel of the image data of the first color component among the predetermined multiple color image data included in the color image data. N rasters are converted into dot rows in parallel while comparing with a predetermined threshold value associated with. For the image data of the second color component, N rasters are converted into dot rows in parallel while diffusing gradation errors caused by determining the presence or absence of dot formation to surrounding undetermined pixels. I will do it.

  As described above, image data often has an approximate gradation value associated with a neighboring pixel. Therefore, if N rasters are converted into dot rows in parallel in this way, cache data is cached. The hit rate, the efficiency of pipeline processing, and the hit prediction accuracy are improved. As a result, image data can be converted quickly. If both the image data of the first color component and the image data of the second color component are processed in parallel for each of N rasters, the image data of these color components is converted into a dot row in parallel. It is preferable that the processing is not unnecessarily complicated.

  In such a second image processing apparatus that handles color image data, the image data of the first color component may be converted as follows. A plurality of threshold values are stored in advance in a two-dimensionally arranged threshold value matrix. Then, N rasters adjacent to each other from a plurality of rasters constituting the image data are collected as a raster group, and each pixel included in the raster group is sequentially selected in a predetermined order, whereby the N rasters are selected. The target pixel is selected in parallel from the raster. The gradation value associated with the selected target pixel and the threshold value corresponding to the pixel are read out from the threshold matrix and compared with each other to determine whether or not dots are formed for the target pixel.

  In this way, it is possible to quickly convert the image data of the first color component into data expressed by the presence or absence of dot formation while converting N rasters into dot rows in parallel.

  In such an image processing apparatus, the threshold values of the threshold matrix may be grouped into a plurality of groups of N rows and stored in a predetermined order in which the target pixels are selected in the group.

  If threshold values in the threshold matrix are stored in this way, when the tone value of the image data for the target pixel and the threshold value corresponding to the target pixel are read to determine the presence or absence of dot formation, It may be read out in the order stored. Therefore, it is preferable because the processing can be simplified. Further, if each threshold value in the threshold value matrix is stored in accordance with the order in which the target pixels are selected in this way, even if the CPU that performs image processing does not have a built-in cache or does not perform pipeline processing. The advantage that the processing can be speeded up is also obtained.

  The target pixel may be selected as follows. That is, a plurality of pixels included in a raster group may be combined into a plurality of blocks each having a predetermined number of pixels in the vertical and horizontal directions, and target pixels may be sequentially selected from the respective pixels in the block.

  By doing this, it is possible to sequentially select the target pixels from the pixels constituting the raster group by repeatedly performing the process of selecting the target pixels from each pixel in the block. As a result, the processing can be simplified, which is preferable.

  In the above-described second image processing apparatus, the image data of the second color component may be converted as follows. First, a raster group is generated by collecting the N rasters adjacent to each other. Then, the last raster at the last of the raster group is converted into a dot row, and the gradation error generated in each pixel is diffused to surrounding undecided pixels. Further, the head raster at the head of the raster group adjacent to the last raster is converted into a dot row, and the gradation error generated in each pixel is diffused to surrounding undecided pixels. Further, with respect to the remaining raster obtained by removing the head raster from the raster group, each residual raster is considered in consideration of the gradation error diffused from pixels that belong to the same raster group as the remaining raster and for which dot formation has been determined. By determining the presence or absence of dot formation of the pixel, the remaining raster is converted into a dot row in parallel with the process of converting the leading raster into the dot row. When these gradation errors are diffused, the error diffused to the pixels of the same raster group as the pixel for which the presence / absence of dot formation is determined and the error diffused to the pixels of the different raster group are stored in different error storage units. . Each raster is converted into a dot row using the error thus diffused.

  By doing so, the dither method is applied to the image data of the first color component, and the error diffusion method is applied to the image data of the second color component, and a plurality of rasters constituting the raster group are applied in parallel. Can be converted to a dot row. As a result, color image data can be quickly converted into data for each color in an expression format depending on whether or not dots are formed.

  In such a second image processing apparatus, the image data of the second color component may be converted as follows. First, the image data of the second color component is collected into blocks each having a predetermined number of pixels in the vertical and horizontal directions, and a plurality of blocks are generated from the raster group. Then, it is also possible to convert the N rasters constituting the raster group into dot rows in parallel by determining whether or not dots are formed in units of blocks.

  As will be described in detail later, if processing is performed in units of blocks in this way, it is possible to quickly determine the presence or absence of dot formation for the image data of the second color component, and thus to quickly convert the image data. It becomes possible.

  When the color image data handled by the second image processing apparatus described above is data composed of image data of a plurality of colors including at least cyan, magenta, and yellow, at least the image data of yellow is You may convert as image data of a 1st color component.

  The yellow color image data has a relatively small influence on the image quality even if the data conversion accuracy is somewhat deteriorated. Therefore, if the yellow color image data is converted as the image data of the first color component, the image quality is deteriorated. It is possible to quickly convert the color image data without causing it.

  In such a second image processing apparatus that handles color image data, at least cyan or magenta image data may be converted as image data of the second color component.

  Since cyan or magenta color image data tends to have a relatively easy influence on image quality due to deterioration in data conversion accuracy, if these image data are converted as image data of the second color component, Color image data can be converted without deteriorating the image quality.

  Further, when the color image data includes black image data, the black image data may be converted as the image data of the second color component.

  In black image data, as in the case of cyan or magenta image data, deterioration in conversion accuracy is relatively likely to affect the image quality. Therefore, if the black image data is converted as the image data of the second color component, the color image data can be converted without deteriorating the image quality.

  Further, when the color image data includes image data of a set of shades having substantially the same hue and different densities, the first image data of the set of colors is the first image data. The image data of the color component may be converted, and the other image data of the set of colors may be converted as the image data of the second color component. For example, when the color image data includes image data for a set of colors consisting of a cyan color and a light cyan color, which is a light cyan color, the conversion is performed as follows. That is, when cyan image data is converted as the first color component, light cyan image data is converted as the second color component, and conversely, cyan image data is converted as the second color component. When converted, the light cyan image data is converted as the first color component. Similarly, when the color image data includes image data for a set of light magenta colors, which are magenta and light magenta, the other is used as image data of the first color component. If converted, the other is converted as image data of the second color component.

  In general, when color image data includes a set of color image data having substantially the same hue and different densities, the magnitudes of the effects of these shaded image data on the image quality are different from each other. When the influence on the image quality is large, the other is in a relation that the influence on the image quality is small. Experience shows that when the density difference between these sets of colors is large, the image data of dark colors has a large effect on image quality. Conversely, when the density differences between sets of colors are small, the image data of pale colors It is known that the influence on the image quality increases. Accordingly, image data having a smaller influence on image quality is converted as image data of the first color component, and image data having a larger influence on image quality is converted into the second image data. If the image data is converted, the color image data can be quickly converted without deteriorating the image quality.

  In addition, both the first image processing apparatus and the second image processing apparatus described above can be suitably applied to a printing apparatus that forms an image by forming dots on a print medium.

  That is, in the first image processing apparatus or the second image processing apparatus described above, image data in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally is received, and the image data is included in the image data. By performing predetermined image processing, data is converted into an expression format depending on the presence or absence of dot formation. In conversion, the image data can be quickly converted by converting a plurality of rasters into dot rows in parallel. Therefore, if image processing is performed on the image data by applying the first image processing apparatus or the second image processing apparatus, the image data can be quickly converted into data in an expression format based on the presence or absence of dot formation. As a result, the image can be printed quickly.

The present invention can also be realized using a computer by causing a computer to read a program that realizes the first image processing method or the second image processing method described above. Therefore, the present invention includes the following aspects as a computer program. That is, the program corresponding to the first image processing method described above is
Receiving image data in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally, comparing the gradation value of each pixel with a predetermined threshold value and determining whether or not a dot is formed A program for realizing, using a computer, a method for converting the image data into data of an expression format depending on whether or not dots are formed by converting a raster as a row of dots into a dot row,
A first function for storing in advance a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged;
A second function of selecting a target pixel for determining whether or not to form a dot;
A third function for reading out a threshold value corresponding to the target pixel from the threshold value matrix and comparing it with the gradation value corresponding to the target pixel, thereby determining whether or not dots are formed for the target pixel. And
The second function is to combine predetermined N adjacent rasters (N is an integer of 2 or more) as a raster group, and sequentially select each pixel included in the raster group according to a predetermined order, The gist of the present invention is to select the target pixel in parallel from the N rasters.

A program corresponding to the second image processing method described above is
Receiving color image data in which gradation values of a plurality of predetermined colors are associated with each of a plurality of pixels arranged two-dimensionally, and determining the presence or absence of dot formation for each pixel for each color component, A program for realizing, by using a computer, a method for converting color image data into data of each color in an expression format depending on the presence or absence of dot formation by converting a raster that is a row into a dot row,
A function of separating the received color image data into image data of a first color component preselected from the color components and image data of a second color component excluding the first color component When,
For the image data of the first color component, a predetermined number of adjacent rasters (N is 2 or more) are compared while comparing the gradation value of each pixel with a predetermined threshold value associated with each pixel. A function that converts the data into an expression format based on the presence or absence of dot formation,
As for the image data of the second color component, the gradation error generated in the pixel determined by determining the presence / absence of dot formation is diffused to the surrounding undetermined pixels and diffused from the surrounding determined pixels. By converting the adjacent rasters into dot rows in parallel while determining the presence / absence of dot formation so as to eliminate the tone error, the data is converted into data in an expression format according to the presence / absence of dot formation. The gist is to provide a function.

  If these programs are read by a computer and the various functions described above are realized by using the computer, image data in which gradation values are associated with a plurality of pixels arranged two-dimensionally is determined depending on whether or not dots are formed. It becomes possible to quickly convert the data into an expression format.

  Furthermore, these programs can be recorded on a recording medium so as to be readable by a computer, and the programs on the recording medium can be read to be realized using a computer. Therefore, the present invention can take the form of a recording medium in which the above-described program is recorded so as to be readable by a computer.

[Other aspects of the invention]
The present invention can also be understood as the following aspects. That is,
Receiving image data in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally, comparing the gradation value of each pixel with a predetermined threshold value and determining whether or not a dot is formed An image processing apparatus that converts the image data into data in an expression format depending on whether or not dots are formed by converting a raster as a row of dots into a dot row,
Threshold matrix storage means for storing a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged;
Target pixel selection means for selecting a target pixel for determining whether or not to form a dot;
A dot formation determination unit that reads out a threshold value corresponding to the target pixel from the threshold value matrix and compares the threshold value with the gradation value corresponding to the target pixel, thereby determining whether or not the target pixel is formed. And
The target pixel selection means collects a predetermined N number of adjacent rasters (N is an integer of 2 or more) as a raster group, and sequentially selects each pixel included in the raster group according to a predetermined order, Means for selecting the target pixel in parallel from each of the N rasters;
The dot formation determination means can be grasped as an image processing apparatus that is a means for determining the presence or absence of dot formation for the target pixel while performing pipeline processing.

Or it is also possible to grasp | ascertain this invention as the following aspects. That is,
Receiving image data in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally, comparing the gradation value of each pixel with a predetermined threshold value and determining whether or not a dot is formed An image processing apparatus that converts the image data into data in an expression format depending on whether or not dots are formed by converting a raster as a row of dots into a dot row,
Threshold matrix storage means for storing a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged;
Target pixel selection means for selecting a target pixel for determining whether or not to form a dot;
A dot formation determination unit that reads out a threshold value corresponding to the target pixel from the threshold value matrix and compares the threshold value with the gradation value corresponding to the target pixel, thereby determining whether or not the target pixel is formed. And
The target pixel selection means collects a predetermined N number of adjacent rasters (N is an integer of 2 or more) as a raster group, and sequentially selects each pixel included in the raster group according to a predetermined order, Means for selecting the target pixel in parallel from each of the N rasters;
The dot formation determination means temporarily stores intermediate information generated when determining the presence or absence of dot formation for the target pixel in a cache memory that can be read and written at a higher speed than the main memory in which an instruction for the determination is stored. It can be grasped as an image processing apparatus which is a means for judging the presence or absence of the dot formation while accumulating it.

In order to more clearly describe the operation and effect of the present invention, embodiments of the present invention will be described below in the following order.
A. Summary of the invention:
B. First embodiment:
B-1. Device configuration:
B-2. Overview of image data conversion process:
B-3. Tone number conversion processing of the first embodiment:
B-4. Variations:
C. Second embodiment:
C-1. Storage method of threshold in gradation number conversion process of the second embodiment:
C-2. Tone number conversion processing of the second embodiment:
C-3. Variations:
D. Third embodiment:
D-1. Tone number conversion processing of the third embodiment:
D-2. Variations:

A. Summary of the invention:
The outline of the present invention will be described with reference to FIG. FIG. 1 is an explanatory diagram conceptually showing an outline of the present invention, taking a printing system as an example. The printing system includes an image processing apparatus 10 according to the present invention and a printer 20. The image processing apparatus 10 receives the image data, performs predetermined image processing, converts it into print data in an expression format depending on whether dots are formed, and outputs the print data to the printer 20. In the drawing, an outline of processing in which the image processing apparatus 10 performs image processing on image data and converts the data is shown in the form of a flowchart. The image data supplied to the image processing apparatus 10 is expressed in a format in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally. Such image data can be generated by performing predetermined conversion on an image taken by an image device such as a digital camera or a color scanner, or an image created on a computer using various application programs.

  When receiving such image data, the image processing apparatus 10 performs image processing roughly in the following flow. First, a pixel (symmetric pixel) to be used to determine whether or not dots are formed is selected from the image data (step S10), and the gradation value associated with the symmetric pixel is associated with the target pixel by a predetermined method. The attached threshold value is read (step S20). As the threshold value, a plurality of threshold values are stored in advance in a matrix. In this specification, such a matrix storing threshold values is called a threshold matrix. In FIG. 1, a threshold value matrix in which a total of 36 threshold values, each of 6 in the vertical and horizontal directions are stored as an example, is displayed. Of course, a larger number of threshold values can be stored in the threshold value matrix. . In step S20, when the threshold value matrix is superimposed on the image data, the threshold value stored at the position corresponding to the symmetric pixel is read as the threshold value associated with the target pixel.

  In FIG. 1, a frame labeled “Symmetric pixel selection” conceptually shows a state where a threshold matrix is superimposed on image data. A thin broken line in the display indicates a plurality of pixels arranged two-dimensionally, and a thick broken line indicates a threshold matrix. Usually, since the size of the threshold matrix is smaller than the size of the image data, the threshold matrix is repeatedly used while moving the position. When the threshold value is read out in this manner, the presence / absence of dot formation for the target pixel is determined by comparing with the gradation value of the target pixel (step S30). Next, it is determined whether or not dot formation has been determined for all the pixels constituting the image data (step S40), and the above-described series of processing is repeated until the determination is completed for all the pixels.

  Here, in the image processing apparatus 10 of the present invention, target pixels are sequentially selected by a predetermined method in order to quickly convert image data. In the figure, “target pixel selection” is displayed in a frame indicating the selection of the target pixel. For convenience of explanation, a pixel row in a plurality of pixels arranged two-dimensionally is called a raster, and image data is considered to be composed of a plurality of rasters. These rasters are numbered, and “raster 1”, “raster 2”,..., And their respective rasters are identified.

  When selecting a target pixel, first, a predetermined number of adjacent rasters are gathered to form a raster group. In the example shown in the figure, a raster group is formed by combining two rasters such as raster 1 and raster 2, raster 3 and raster 4, or raster 5 and raster 6. That is, in the illustrated example, one raster group includes a plurality of pixels corresponding to two rasters. Next, the target pixels are sequentially selected from the plurality of pixels constituting the raster group in a predetermined order. To explain according to the example in the figure, first, the leftmost pixel of the raster 1 is selected as the target pixel, and then the leftmost pixel of the raster 2 is selected. Next, the second pixel from the left of raster 1 is selected, and then the second pixel from the left of raster 2 is selected. Thus, the target pixel is alternately selected from the raster 1 and the raster 2. In the drawing, a solid line that proceeds in a zigzag shape from the left end toward the right direction on the raster 1 and the raster 2 schematically shows how the target pixel is alternately selected from these two rasters. In this way, by determining the presence or absence of dot formation while alternately selecting target pixels from the two rasters, raster 1 and raster 2 are converted into dot rows in parallel.

  When the presence or absence of dot formation is determined for all the pixels of raster 1 and raster 2 as described above, the target pixel is selected from the pixels of raster 3 and raster 4 this time. That is, the leftmost pixel of raster 3 is selected as the target pixel, then the leftmost pixel of raster 4, the second pixel from the left of raster 3, the second pixel from the left of raster 4, and so on. Four pixels are alternately selected as target pixels. In the figure, a thick dashed arrow heading from the right side to the leftmost pixel of the raster 3 just below the raster 2 indicates that the processing of the raster 1 and the raster 2 has been completed and the processing has shifted to the processing of the raster 3 and the raster 4. This is shown schematically. Then, if the presence or absence of dot formation is determined for all the pixels while alternately selecting the raster 3 and raster 4 pixels, the target pixel is selected from the raster 5 and raster 6 pixels. When the processing of the raster 5 and the raster 6 is completed, a plurality of rasters are converted into dot rows in parallel for each raster group, such as the raster 7 and the raster 8.

  In the above description, each raster group has been described as being composed of two rasters. However, it is also possible to construct a larger number of rasters. Further, the order in which each pixel constituting the raster group is sequentially selected as the target pixel is not limited to that described above, and may be an order in which the rasters constituting the raster group are converted into dot rows in parallel. For example, various orders can be used.

  As described above, if a plurality of rasters constituting a raster group are converted into dot rows in parallel, they can be quickly converted for various reasons as follows.

  In general, image data tends to be associated with approximate gradation values between neighboring pixels. When determining the presence or absence of dot formation continuously for a plurality of pixels, if these pixels have gradation values that are close to each other, the execution code generated internally by the CPU to make the determination is stored in the cache memory. The possibility of remaining is high. If the execution code remains, the CPU can read the execution code from the cache memory and proceed with the process, so that it is not necessary to generate the execution code, and the process can be speeded up accordingly. The probability that necessary execution code or data or the like remains in the cache memory may be referred to as a cache hit rate. Here, since the pixels are two-dimensionally arranged, there is a high possibility that similar approximate gradation values are associated with pixels that are adjacent in either the vertical or horizontal diagonal directions. For example, the pixel at the left end of the raster 1 shown in FIG. 1 will be described as an example. Pixels around this pixel, that is, the second pixel from the left end of the raster 1, the second pixel from the left end or the left end of the raster 2 In any case, there is a high possibility that a gradation value approximated to the pixel at the left end of the raster 1 is associated. Therefore, if the target pixel is alternately selected from the pixels of raster 1 and raster 2 to determine the presence or absence of dot formation, pixels having approximate gradation values are successively processed from the two-dimensionally spread pixels. Therefore, it is possible to improve the cache hit rate by that amount and speed up the processing.

  In addition, when the CPU that performs image processing is performing pipeline processing or branch prediction of pipeline processing, it can be further accelerated by processing multiple rasters in parallel for the following reasons. Can be realized. When pixels with similar gradation values are processed continuously, there is a high possibility that similar processing is repeated. For example, when the gradation values of pixels that are continuously processed are sufficiently low, there is a high possibility that it is determined that no dot is formed for any pixel. That is, while these pixels are processed, there is a high possibility that similar processing is continuously performed inside the CPU. For this reason, the efficiency of pipeline processing and the hit rate of branch prediction are improved, and as a result, the processing can be speeded up. As described above, if a plurality of rasters are processed in parallel, pixels that are close to each other in the vertical and horizontal directions can be processed in succession. Can be processed in succession. As a result, the efficiency of pipeline processing and the hit rate of branch prediction can be improved, and the processing can be speeded up accordingly.

  Various aspects exist in the image processing apparatus of the present invention. Below, these various aspects are demonstrated in detail using an Example.

B. First embodiment:
B-1. Device configuration:
FIG. 2 is an explanatory diagram illustrating a configuration of the computer 100 as the image processing apparatus according to the first embodiment. The computer 100 is a well-known computer configured by connecting a ROM 104, a RAM 106, and the like with a bus 116 around a CPU 102. The CPU 102 has a built-in cache memory. When reading information from the ROM 104 or the RAM 106, the CPU 102 first accesses the cache memory, and if necessary information remains, reads it from the cache memory. We are trying to speed up. Further, in the CPU 102, the processing speed is increased by a parallel operation method called so-called pipeline processing, and branch prediction control is also performed in order to make the pipeline processing function effectively.

  The computer 100 includes a disk controller DDC 109 for reading data on the flexible disk 124 and the compact disk 126, a peripheral device interface P-I / F 108 for exchanging data with peripheral devices, and a video interface for driving the CRT 114. A V-I / F 112 or the like is connected. A color printer 200 (to be described later), a hard disk 118, and the like are connected to the P-I / F 108. Further, if a digital camera 120, a color scanner 122, or the like is connected to the P-I / F 108, an image captured by the digital camera 120 or the color scanner 122 can be printed. If the network interface card NIC 110 is attached, the computer 100 can be connected to the communication line 300 to acquire data stored in the storage device 310 connected to the communication line.

  FIG. 3 is an explanatory diagram showing a schematic configuration of a color printer 200 used in combination with the image processing apparatus of the first embodiment. The color printer 200 is an ink jet printer capable of forming dots of four color inks of cyan, magenta, yellow, and black. Of course, in addition to these four color inks, an inkjet printer capable of forming a total of six ink dots including cyan (light cyan) ink with a low dye density and magenta (light magenta) ink with a low dye density is used. You can also. Furthermore, an ink jet printer that can form ink dots of a total of seven colors including low (light) yellow (dark yellow) ink may be used. In the following, cyan ink, magenta ink, yellow ink, black ink, light cyan ink, light magenta ink, and dark yellow ink are respectively changed to C ink, M ink, Y ink, K ink, LC ink, and LM. Ink and DY ink are abbreviated.

  As shown in the figure, the color printer 200 drives a print head 241 mounted on a carriage 240 to eject ink and form dots, and the carriage 240 is reciprocated in the axial direction of a platen 236 by a carriage motor 230. And a control circuit 260 that controls dot formation, carriage 240 movement, and printing sheet conveyance.

  An ink cartridge 242 that stores K ink and an ink cartridge 243 that stores various inks of C ink, M ink, and Y ink are mounted on the carriage 240. When the ink cartridges 242 and 243 are mounted on the carriage 240, each ink in the cartridge is supplied to ink discharge heads 244 to 247 for each color provided on the bottom surface of the print head 241 through an introduction pipe (not shown). FIG. 4 is an explanatory diagram showing a state in which a plurality of nozzles Nz are provided on the bottom surface of the print head 241. As shown in the drawing, a plurality of nozzles Nz are arranged in two rows in a staggered pattern at a fixed nozzle pitch k for each color ink discharge head on the bottom surface of the print head 241.

  The control circuit 260 shown in FIG. 3 includes a CPU 261, a ROM 262, a RAM 263, and the like, and controls main scanning and sub-scanning of the carriage 240 by controlling operations of the carriage motor 230 and the paper feed motor 235. Based on the print data supplied from the computer 100, the ink ejection head for each color is driven to eject ink droplets at an appropriate timing. Thus, the color printer 200 can print a color image by forming ink dots of respective colors at appropriate positions on the print medium under the control of the control circuit 260.

  Various methods can be applied to the method of ejecting ink droplets by driving the ink ejection head of each color. That is, a method of ejecting ink using a piezoelectric element, a method of ejecting ink droplets by generating bubbles in the ink passage with a heater arranged in the ink passage, and the like can be used. Also, instead of ejecting ink, use a method that uses ink transfer to form ink dots on printing paper using a phenomenon such as thermal transfer, or a method that uses static electricity to attach toner powder of each color onto the print medium. It is also possible to do.

  In addition, the size of the ink droplets ejected from the ink ejection head, or the method of controlling the number of ink droplets ejected by ejecting a plurality of fine ink droplets on the printing paper is used. It is also possible to use a printer capable of controlling the size of ink dots to be formed, a so-called variable dot printer.

B-2. Overview of image data conversion process:
FIG. 5 is a flowchart showing a flow of processing in which the computer 100 as the image processing apparatus of the first embodiment converts the received image data into print data by applying predetermined image processing to the received image data. Such processing is started when the operating system of the computer 100 activates the printer driver, and is executed using the function of the CPU 102. The image data conversion process of the first embodiment will be briefly described below with reference to FIG.

  When starting the image data conversion process, the printer driver first starts reading color image data (step S100). The color image data read here is a so-called RGB image in which gradation values for each color of red (R), green (G), and blue (B) are associated with a plurality of pixels arranged two-dimensionally. It is data.

  Next, the resolution of the captured image data is converted to a resolution for printing by the color printer 200 (step S102). When the resolution of color image data is lower than the printing resolution, new data is generated between adjacent image data by performing an interpolation operation. Conversely, when the resolution is higher than the printing resolution, the data is thinned out at a certain rate. To convert the resolution of the image data to the printing resolution. Further, for convenience, instead of performing the interpolation calculation, a pixel having the same gradation value as any of the pixels can be newly generated between adjacent pixels.

  When the resolution is converted in this way, color conversion processing of color image data is performed (step S104). Color conversion processing refers to RGB image data expressed by a combination of R, G, and B gradation values, by a combination of gradation values of each color used in the color printer 200 such as C, M, Y, and K. This is a process of converting into expressed image data. The color conversion process can be quickly performed by referring to a three-dimensional numerical table called a color conversion table (LUT).

  When the color conversion process is finished, the gradation number conversion process is started (step S106). The gradation number conversion process is the following process. The image data obtained by the color conversion process is data having a gradation width of 256 gradations for each color of C, M, Y, and K. On the other hand, the color printer 200 of the first embodiment can take only the state of “form dots” or “do not form dots”. That is, the color printer 200 of the present embodiment can only express two gradations locally. Therefore, it is necessary to convert image data having 256 gradations into image data expressed in 2 gradations that can be expressed by the color printer 200. A process for converting the number of gradations is a gradation number conversion process.

  In the gradation number conversion processing of the first embodiment, processing is performed using a so-called dither method. In the dither method, the gradation value of a pixel is compared with a threshold value associated with the pixel, and if the gradation value is larger, it is determined that a dot is formed in that pixel. This is a method of converting image data into data in an expression format based on the presence or absence of dot formation by determining that no dot is formed if it is small. In the dither method, a plurality of threshold values associated with pixels are stored in a matrix arrangement, and in accordance with the position in the image data of the pixel for which it is determined whether or not to form dots, Appropriate thresholds are associated. In this specification, such a matrix in which a plurality of threshold values are recorded is called a threshold value matrix.

  FIG. 6A is an explanatory diagram illustrating a state in which a part of the threshold matrix is enlarged and a plurality of threshold values are stored. In the example threshold matrix, thresholds corresponding to a total of 4096 pixels are stored in a square matrix of 64 pixels vertically and horizontally. Of course, the size of the threshold matrix is not limited to this, and may be various sizes as required. Further, the matrix is not necessarily limited to a square matrix, and may be a matrix having different numbers of vertical and horizontal pixels, for example. Note that the use of a large threshold matrix as shown in FIG. 6A makes the description complicated, and therefore the following description assumes that the threshold matrix is a small matrix as shown in FIG. 6B. To do.

  FIG. 7 is an explanatory diagram conceptually showing a state in which a threshold value matrix is applied to image data to determine the presence or absence of dot formation for each pixel. Each rectangle indicated by a thin broken line in FIG. 7 represents a pixel. The image data is expressed in a format in which a gradation value is associated with each of a plurality of pixels arranged two-dimensionally. In order to simplify the description, it is assumed here that gradation values 64 are associated with all the pixels. FIG. 7A schematically shows how a threshold matrix is applied to such image data. In FIG. 7A, the threshold matrix is represented by a thick solid line. As described above, threshold values are stored in advance in each cell of the threshold value matrix. As shown in FIG. 7A, by superimposing the threshold value matrix on the image data, each pixel constituting the image data can be associated with each square of the threshold value matrix.

  In the dither method, after associating the pixels of the image data with the squares of the threshold matrix in this way, the gradation value of each pixel is compared with the threshold value stored in each square. Then, it is determined that a dot is formed in a pixel having a larger gradation value of the pixel, and a dot is not formed in a pixel that is not so. Here, since it is assumed that gradation values 64 are stored in all the pixels, it is determined that “dots are formed” for pixels having a threshold value of 63 or less stored in the threshold value matrix, and the threshold value is 64 or more. For these pixels, it is determined that “no dot is formed”. In FIG. 7A, the pixels determined to form dots are displayed with hatching.

  Since the image data is composed of a large number of pixels, it is usually larger than the threshold matrix. Therefore, the same matrix is repeatedly used while moving the position of the threshold matrix with respect to the image data in order to determine the presence or absence of dot formation for all pixels. FIG. 7B shows the result of determining whether or not dots are formed for all the pixels. By determining whether or not dots are formed as described above, dots can be formed in an appropriate ratio according to the gradation value of the image data and in a state where the dots are dispersed from each other. For example, here, the gradation value of the image data is assumed to be 64, which corresponds to about ¼ of the full-scale gradation value 256. Correspondingly, in FIG. 7B, dots are formed in about 1/4 of all the pixels. The density at which dots are formed increases as the gradation value of the image data increases, and decreases as the gradation value decreases. Thus, in the dither method, by setting an appropriate threshold value in the threshold matrix, dots can be formed with an appropriate density according to the gradation value of the image data.

  In the above description, it is determined whether or not dots are formed for all the pixels in the threshold matrix, and then the position of the threshold matrix is shifted to associate the matrix with new pixels, and whether or not dots are formed for these pixels is determined. Explained as if. However, this is for the convenience of explanation. Actually, the presence / absence of dot formation is determined for each raster rather than for each threshold matrix. That is, the presence or absence of dot formation is determined while proceeding in the right direction from the left end on one raster, and when the determination is completed for all pixels on the raster, processing of the lower raster is started. In this way, in practice, the presence / absence of dot formation is determined for each raster. On the other hand, in the tone number conversion process of the present embodiment, it is possible to perform a quick process by determining in parallel whether or not a plurality of raster dots are formed. A method of determining whether or not dots are formed on a plurality of rasters in parallel will be described in detail later.

  In the image data conversion process shown in FIG. 5, after the gradation number conversion process is completed as described above, the interlace process is started (step S108 in FIG. 5). The interlace process is a process of rearranging the image data converted into the expression format depending on the presence or absence of dot formation into the order to be transferred to the color printer 200 in consideration of the dot formation order. The printer driver finally outputs the image data obtained by performing the interlace processing to the color printer 200 as print data (step S110). The color printer 200 forms ink dots of each color on the print medium according to the print data thus supplied. As a result, a color image corresponding to the image data is printed on the print medium.

  Hereinafter, in the tone number conversion process of the first embodiment, a process for quickly converting the tone number of image data by determining whether or not dots are formed in parallel for a plurality of rasters will be described.

B-3. Tone number conversion processing of the first embodiment:
FIG. 8 is a flowchart showing the flow of the tone number conversion process of the first embodiment. This process is a process performed by the printer driver of the computer 100 in the image data conversion process shown in FIG. The processing shown in FIG. 8 is performed for each color of C, M, Y, and K. However, in order to avoid complicated description, the following description will be made without particularly specifying a color. FIG. 9 is an explanatory diagram conceptually showing how the printer driver determines the presence / absence of dot formation according to the tone number conversion process of the first embodiment. Hereinafter, a description will be given according to the flowchart of FIG. 8 with reference to FIG.

  When starting the tone number conversion process, the printer driver first performs a process of forming a raster group by combining a predetermined number of adjacent rasters (step S200). Here, it is assumed that a raster group is formed by collecting two rasters. Of course, the number of rasters is not limited to two, and a larger number of rasters may be collected as a raster group. In FIG. 9, “raster 1” at the top of the image and “raster 2” below it are hatched to indicate that these rasters are grouped as a raster group.

  Next, one target pixel is selected from among a plurality of pixels included in the raster group to determine whether or not dots are formed (step S202). Initially, the pixel at the upper left of the image data, that is, the leftmost pixel of the raster 1 is selected as the target pixel. In FIG. 9, this pixel is indicated by a number [1].

  Thereafter, the gradation value Dt of the selected target pixel is read (step S204), and then the threshold value th corresponding to the target pixel is read from the threshold matrix (step S206). Here, as described above with reference to FIG. 7A, the threshold value corresponding to the target pixel is a matrix of the threshold matrix at a position that overlaps the target pixel when the threshold matrix is superimposed on the image data. The threshold value stored in the eye. A rectangle indicated by a thick broken line in FIG. 9 schematically shows a state in which a threshold matrix is superimposed on image data.

  When the gradation value Dt of the target pixel and the corresponding threshold value th are read out in this way, the magnitude relationship between these is compared (step S208). If the gradation value Dt is larger (step S208: yes), it means that it is determined that a dot is to be formed on the target pixel, and that a dot is formed on the variable Dr indicating the determination result of the presence or absence of dot formation. The value “1” to be written is written (step S210). On the contrary, when the threshold th is larger, a process of writing a value “0” indicating that no dot is formed is performed in the variable Dr indicating the determination result (step S212).

  When it is determined whether or not dots are formed for the selected target pixel as described above, it is determined whether or not all the pixels included in the raster group are selected as the target pixels (step S214). Here, only the pixel numbered “1” in FIG. 9 has been processed, and many unprocessed pixels remain in the raster group. Therefore, it is determined as “no” in the process of step S214, and the process returns to step S202 to select a new target pixel.

The new target pixel is performed according to a predetermined rule so that the target pixel is selected sequentially from each raster included in the raster group. Here, it is assumed that the target pixel is selected according to the following rule.
<< Rule 1 >>: When the target pixel is on an odd-numbered raster, a pixel immediately below the target pixel is selected as a new target pixel. That is, the raster number is incremented by one in the left-right direction without changing the pixel position.
<< Rule 2 >>: When the target pixel is on the even-numbered raster, the pixel immediately above the target pixel is selected as a new target pixel. That is, the raster number is decremented by one and the number indicating the pixel position is incremented by one.

  As a result of selecting a new target pixel according to such a rule, a pixel numbered [2] in the figure is selected as a new target pixel next to the pixel numbered [1] in FIG. It will be. The pixel of the number [2] selected in this way is subjected to the processing from step S204 to step S214 as described above, and then returns to step S202 again to select the next target pixel. According to the rule described above, the pixel with the number [3] is selected as a new target pixel after the pixel with the number [2]. In the same manner, the target pixel is sequentially selected in the order of the pixel with the number [4], the pixel with the number [5], and the pixel with the number [6], and whether or not dots are formed is determined for each pixel. To go.

  Further, as shown in FIG. 9, when the determination is made for the pixel of the number [12], the right end of the threshold matrix is reached. The threshold matrix may be moved in the right direction. The arrows shown in FIG. 9 schematically represent how the threshold matrix is moved.

  As described above, whether or not dots are formed is determined while sequentially selecting target pixels from a plurality of pixels included in the raster group in a predetermined order. In this way, it is possible to determine whether or not dots are formed while the raster 1 and the raster 2 are parallel.

  When it is determined whether or not dots are formed for all the pixels included in the raster 1 and the raster 2, that is, when “yes” is determined in step S214 in FIG. 8, processing of all the rasters included in the image data is performed. It is determined whether or not the processing has been completed (step S216). Here, only two rasters, raster 1 and raster 2, are processed, and a large number of rasters still remain unprocessed. Therefore, it is determined as “no” in step S216, and the process returns to step S200. 3 and raster 4 are processed to generate a new raster group. For the new raster group generated in this way, by selecting the target pixel in the same manner, it is determined whether or not dots are formed while the raster 3 and the raster 4 are parallel. When the determination for all the pixels for the raster 3 and the raster 4 is completed, the raster 5 and the raster 6 are generated as a new raster group, and the presence or absence of dot formation is determined for these rasters. A series of processes as described above are repeated until all the rasters included in the image data have been processed. When the processing of all rasters is completed (step S216 in FIG. 8: yes), the process returns to the image data conversion process in FIG. 5 after exiting the tone number conversion process in the first embodiment shown in FIG.

  As described above, according to the tone number conversion process of the first embodiment, the presence or absence of dot formation can be determined in parallel for two rasters collected as a raster group. As described above, adjacent pixels in the image data tend to be associated with tone values that are close to each other. Thus, if multiple rasters are processed in parallel, the cache hit rate, pipeline processing, etc. As a result, the accuracy of branching or the accuracy of branch prediction is improved, so that the gradation number conversion process can be performed quickly. As a result, the image data can be promptly converted into data in an expression format depending on whether or not dots are formed, so that the image can be printed quickly.

B-4. Variations:
In the first embodiment described above, two rasters are grouped together as a raster group, but a larger number of rasters may be grouped as a raster group instead of two.

FIG. 10A shows a state where three rasters are grouped as a raster group as an example. In the drawing, a rectangle indicated by a thick broken line schematically shows a state in which three rasters are grouped as a raster group. Even when the three rasters are combined as a raster group in this way, the target pixels may be sequentially selected from these rasters according to a predetermined rule. In this way, it is possible to determine whether or not dots are formed by paralleling the plurality of rasters. For example, paying attention to the remainder when the number of the target pixel is divided by 3, the target pixel can be selected based on the following rules.
<< Rule 3 >>: If the number of a raster with a target pixel is not divisible by 3, the raster number is incremented by one without changing the pixel position in the left-right direction.
<< Rule 4 >>: When the number of a raster with a target pixel is divisible by 3, the raster number is decremented by two and the number indicating the pixel position is incremented by one.
If the target pixels are sequentially selected based on such a rule, as indicated by the arrows in FIG. 10A, the pixel with the number [1], the pixel with the number [2], and the number [3 ] And the pixel number [4] are selected in order, and the presence or absence of dot formation can be sequentially determined for these pixels.

  Further, the method for selecting the target pixel from the plurality of pixels constituting the raster group is not limited to the above-described method, and various methods can be applied. For example, as indicated by arrows in FIG. 10B, the target pixels may be sequentially selected so as to cross the raster group in an oblique direction.

  Alternatively, the raster group may be divided into small blocks, and the pixels constituting each block may be sequentially selected in a predetermined order. For example, in the example shown in FIG. 10C, a raster group consisting of two rasters is divided into small rectangular blocks each consisting of two pixels in the vertical and horizontal directions and a total of four pixels, and the pixels in each block are arranged in a predetermined order. select. Referring to FIG. 10C, the pixel numbered [1] is selected first, and then the pixel numbered [2] on the right is selected. Next, the pixel numbered [3] directly below the pixel numbered [1] is selected, and finally, the pixel numbered [4] next to the right is selected. Thus, the pixels in the block may be selected in order for each block.

  Further, if the pixels in each block are selected in this way, the pixels in the block can be sequentially selected almost corresponding to the distance from the first selected number [1] in the block. The closer the distance between the pixels, the more likely that the approximate gradation value is associated. Therefore, if the pixels are sequentially selected so as to substantially correspond to the distance, the cache hit rate, pipeline processing Efficiency or the accuracy of branch prediction can be improved, and as a result, the number of gradation conversion processing can be speeded up.

  In the example shown in FIG. 10C, first, all the pixels of raster 1 (pixels numbered [1] and [2] in the figure) are selected, and then the pixels of raster 2 (in the figure). All the pixels numbered [3] and [4] are selected. Thus, when the raster group is divided into small blocks, the target pixel may be selected in units of rasters in the block. Even in this way, it is possible to process the rasters constituting the raster group in parallel when viewed as a whole.

C. Second embodiment:
As described above, when a plurality of rasters are grouped into a raster group and the gradation number conversion process is performed while selecting symmetric pixels from each raster in a predetermined order, the threshold matrix of the threshold matrix is considered in consideration of the order of selecting pixels. By storing the threshold value, the processing speed can be increased. Hereinafter, the gradation number conversion process of the second embodiment will be described.

C-1. Storage method of threshold in gradation number conversion process of the second embodiment:
For convenience of explanation, the cells constituting the threshold value matrix are numbered as follows. For example, the threshold value (m, n) indicates the threshold value stored in the m-th row and n-th column of the threshold value matrix. By identifying each cell in this way, a plurality of threshold values stored in the threshold value matrix can be identified. FIG. 11 is an explanatory diagram showing how the threshold values stored in the threshold value matrix are numbered and identified in this way.

  In the gradation number conversion process of the second embodiment, each threshold value of the threshold value matrix is stored in consideration of the order of selecting symmetric pixels from each raster. In the following, the gradation number conversion process of the second embodiment will be described by taking as an example the case of selecting symmetric pixels in the order shown in FIG. As described above, in the example shown in FIG. 9, when two rasters are combined to form a raster group, and the symmetric pixel is on the odd-numbered raster, the pixel immediately below the symmetric pixel is newly symmetric. Select as a pixel (see Rule 1 above), and if the symmetric pixel is on an even-numbered raster, select the pixel immediately diagonally right above the symmetric pixel as a new symmetric pixel (see Rule 2 above) .

  Corresponding to the selection of the symmetric pixels in this order, the threshold values of the threshold value matrix are stored in the following order in the second embodiment. First, the first row and the second row of the threshold matrix are paired. For convenience, this set will be referred to as “unit A” below. This corresponds to forming a raster group by grouping two adjacent rasters together when selecting a symmetric pixel.

  First, the threshold value (1, 1) stored in the first row and first column of the threshold matrix is stored, and then the threshold value (2, 1) stored in the immediately lower cell is stored. Record. This corresponds to the above-described << Rule 1 >> for selecting a symmetric pixel, that is, when the symmetric pixel is on an odd-numbered raster, the pixel immediately below is selected as a new symmetric pixel. ing. On the hardware side, this corresponds to storing the threshold (1, 1) at a predetermined address of the ROM 104 or RAM 106 of the computer 100 and storing the threshold (2, 1) at the next address. To do.

  Next, the threshold value (1, 2) stored in the upper right cell of the second row and first column cell is stored. This corresponds to the above-described << Rule 2 >> when selecting a symmetric pixel, that is, when the symmetric pixel is on an even-numbered raster, the pixel immediately above right is selected as a new symmetric pixel. ing. On the hardware, the threshold (1, 2) is stored at the address next to the address where the threshold (2, 1) is stored.

  After storing the threshold value (2, 1), the threshold value (1, 2) immediately below corresponding to rule 1 is stored again, and then the threshold value (2, 2) is stored corresponding to rule 2. Hereinafter, similarly, the threshold value (1, 3), the threshold value (2, 3), and the threshold value (1, 4) are sequentially stored. Thus, when all the threshold values included in the first and second rows (that is, unit A) of the threshold matrix are stored, this time, the threshold values in the third row and the fourth row are combined into “unit B”. Form. The threshold values included in the unit B are also stored sequentially in the order corresponding to the rules 1 and 2. That is, first, the threshold value (3, 1) stored in the upper left corner of the unit B is stored, and then the threshold value (4, 1) of the lower right cell is stored. Further, the threshold value (3, 2) of the upper right cell is stored immediately, and the threshold value (4, 2) and threshold value (3, 3) are sequentially stored in the same manner. When all the threshold values included in unit B have been stored in this way, the threshold values in the threshold matrix are summarized as “unit C” and the threshold values included in unit C are the same. And memorize sequentially.

  FIG. 12A schematically illustrates how the threshold value matrix is divided into three units each having two rows, and the threshold values are sequentially stored while reflecting the rules 1 and 2 in the unit. FIG. The numbers shown in the brackets in the figure indicate the numbers in which the threshold values are stored. For example, the display [5] shown in the square of the threshold matrix indicates that the threshold of the square is stored fifth in all the thresholds. FIG. 12B schematically shows how these threshold values are sequentially stored at consecutive addresses on the ROM 104 or RAM 106 of the computer 100.

C-2. Tone number conversion processing of the second embodiment:
In this way, when a plurality of rasters are combined into a raster group and the number of gradations of image data is converted while paralleling each raster, the threshold matrix threshold is set in consideration of the order of selecting symmetric pixels from each raster. By storing, gradation number conversion processing can be performed quickly. Hereinafter, the tone number conversion process of the second embodiment will be described with reference to FIGS. 8 and 9.

  As described above with reference to the flowchart of FIG. 8, also in the tone number conversion process of the second embodiment, first, predetermined adjacent N rasters are first collected as a raster group (corresponding to step S200 of FIG. 8). Then, a symmetric pixel is sequentially selected from a plurality of pixels constituting the raster group in a predetermined order, and the gradation value of the symmetric pixel is compared with a threshold value to determine whether or not to form a dot on the symmetric pixel. This process is repeated (equivalent to step S202 to step S214 in FIG. 8). When it is determined whether or not dots are formed for all pixels included in the raster group (corresponding to step S214: yes), a new raster group is generated (corresponding to step S200), and the same processing is performed for each pixel of the new raster group. repeat. The same process is repeated until the process is completed for all rasters included in the image data.

  As an example, when two adjacent rasters are grouped into a raster group and symmetric pixels are sequentially selected from a plurality of pixels included in the raster group according to the rules 1 and 2, the schematic diagram shown in FIG. Thus, the gradation number conversion process is performed. That is, first, the gradation value and the threshold value are read for the leftmost pixel of the raster 1 (the pixel numbered as [1] in FIG. 9), and dots are formed based on the magnitude relationship between them. It is determined whether or not. Next, the gradation value and the threshold value are read out for the pixel numbered [2] in FIG. 9, and it is determined whether or not dots are formed. In the same manner, for the pixels numbered [3], [4], and [5], the gradation value and the threshold value are read out to determine whether or not dots are formed.

  Here, since the image data is actually stored in the RAM 106 of the computer 100, the pixel number [1], the pixel number [2], the pixel number [3] and the gradation value are read out. For this purpose, the CPU 102 needs to calculate an address every time the gradation value is read, and calculate an address value in which the gradation value of these pixels is stored on the RAM 106. On the other hand, since the threshold value is stored in advance in consideration of the order in which the target pixel is selected, it is not necessary to perform complicated address calculation in order to read the threshold value. For example, if each threshold value of the threshold value matrix is stored as shown in FIG. 12 in consideration that the target pixel is selected in the order shown in FIG. 9, the threshold value of the number [1] in FIG. If the address at which the number is stored is obtained, the threshold values of the pixels from the number [2] to the number [12] can be sequentially read out by simply incrementing the address value one by one thereafter. That is, the threshold value corresponding to each pixel can be read by simply reading the threshold values of the unit A shown in FIG.

  As for the pixel with the code number [13], as described above, the threshold value of the pixel with the code number [1] is used as a result of moving the position of the threshold value matrix in the raster direction with respect to the image data. Therefore, the threshold value of the pixel of the code number [13] may be read from the address where the threshold value for the code number [1] is stored. Thereafter, similarly, the threshold values of the pixels from the number [14] to the number [24] can be sequentially read by sequentially reading out the threshold values of the unit A while incrementing the address value one by one. . The threshold values can be read out for the subsequent pixels in exactly the same manner. As a result, while the raster group consisting of raster 1 and raster 2 is being processed, the threshold value for each pixel is simply read out repeatedly. Can be read out.

  After all the pixels included in raster 1 and raster 2 have been processed, this time, raster 3 and raster 4 are combined into a raster group, the target pixels are sequentially selected, and dot formation is performed while reading out the gradation value and threshold value of each pixel. Judgment is made. Here, as the threshold value of the target pixel, it is only necessary to repeatedly read the threshold value of the unit B of the threshold value matrix described above with reference to FIG. When all the pixels of the raster 3 and the raster 4 are processed, the raster 5 and the raster 6 are collected as a raster group, and the gradation number conversion process is performed for each pixel while reading the threshold value of the unit C of the threshold value matrix this time. . Thus, when the processing for the rasters 5 and 6 is completed, the rasters 7 and 8 are collected as a raster group, and the gradation number conversion processing for each pixel is performed while repeatedly reading out the threshold value of the unit A again. It is sufficient to read out the threshold value repeatedly from each unit of the threshold value matrix in the same manner for the subsequent rasters.

  FIG. 13 is an explanatory diagram schematically showing how the threshold values of the threshold value matrix stored in the ROM 104 or RAM 106 of the computer 100 are repeatedly read for each unit. As shown in the figure, threshold values in the threshold matrix are stored separately for each of unit A, unit B, and unit C, and within each unit, each threshold value selects the target pixel during the gradation number conversion process. Are stored in consecutive addresses in a predetermined order considering the order in which they are to be performed.

  While the printer driver of the computer 100 is performing the tone number conversion processing of the raster groups of raster 1 and raster 2, the threshold value stored in unit A is indicated by the thin solid arrow in FIG. Are used repeatedly in order. When the raster group that performs the tone number conversion process moves to a raster group consisting of the raster 3 and the raster 4, the threshold values stored in the unit B are set in order as indicated by the thin dashed arrows in the figure. When the raster group to be repeatedly read, used, and processed shifts to the raster group consisting of the raster 5 and the raster 6, it is stored in the unit C as indicated by the thin dashed line arrow in the figure. The threshold value is repeatedly read in order and used. Next, when the raster group to be processed moves to the raster group consisting of the raster 7 and the raster 8, as shown by the thick solid line arrow in the figure, the head address value is returned again and stored in the unit A. The read threshold values are repeatedly read in order and used. Thereafter, the threshold value in the unit may be read in the same manner until the tone number conversion process is completed for all the pixels included in the image data.

  As described above, in the tone number conversion process of the second embodiment, the threshold values of the threshold value matrix are stored in an appropriate order in consideration of the order of selecting the target pixels. In this way, when determining the presence or absence of dot formation, it is possible to determine the presence or absence of dot formation by reading the threshold without performing complicated address calculation for reading the threshold corresponding to the target pixel. Therefore, in the gradation number conversion process of the second embodiment, it is possible to perform a quick conversion.

  In the above description, when the target pixels are selected in the order shown in FIG. 9, that is, two rasters are combined to form a raster group, and according to the above-described << Rule 1 >> and << Rule 2 >>. The case where the target pixels are sequentially selected has been described as an example. Of course, if the number of rasters constituting the raster group and the order in which the target pixels are selected are different, the order in which the threshold values are stored in the threshold value matrix will be different accordingly.

  For example, if the order of selecting the target pixel is as shown in FIG. 10B, the threshold values in the threshold matrix may be stored in the order shown in FIG. That is, the threshold value matrix is grouped into two units of three rows in correspondence with the fact that a raster group is formed by three rasters in FIG. Then, the threshold values may be stored in an order that crosses the unit diagonally according to the order in which the target pixels are selected. Here, the numbers shown in square brackets in FIG. 14 indicate the order in which the threshold values are stored. When the target pixels are selected in the order shown in FIG. 10C, the threshold matrix is grouped into units of two rows and the target pixels are selected as shown in FIG. 14B. Each threshold value may be stored in accordance with.

  Further, in the above description, the case where the order of selecting the target pixel and the order of storing the respective threshold values are completely matched is described, but these orders do not necessarily need to be completely matched, If there is some difference, there may be a deviation in order. For example, when the order in which the target pixel is selected is as shown in FIG. 10C, the threshold values in the threshold matrix may be stored in the order shown in FIG. As is apparent from a comparison between FIG. 10C and FIG. 12A, the order of selecting the target pixel and the order of storing the threshold values are the number [2] pixel and the number [3]. The order is changed with the pixels. Similarly, the pixel number [6] and the pixel number [7] are switched in order. As described above, the order in which the target pixel is selected and the order in which the threshold value is stored are switched between the pixel in which the remainder is 2 when the number of the number is divided by 4 and the pixel in which the remainder is 3. The other pixels are the same. In such a case, it is possible to read out the threshold values without performing complicated address calculation even though the order in which the threshold values are stored in the threshold value matrix does not completely match the order in which the target pixels are selected. is there. For example, if the remainder is 2 when the order of the target pixel is divided by 4, the threshold of the next order is read, and if the remainder is 3 when the order of the target pixel is divided by 4, By reading the previous threshold value, the threshold value corresponding to the target pixel can be easily read. Alternatively, instead of sequentially reading out the threshold values one by one, for example, four threshold values may be read together and an appropriate threshold value selected from these four threshold values. In any case, the order in which the threshold values are stored in the threshold matrix may be an order that considers the order in which the target pixels are selected, and the order in which the threshold values are stored and the order in which the target pixels are selected are completely the same. There is no need to be.

  As described above, in the tone number conversion process of the second embodiment, the threshold values in the threshold matrix are stored in an appropriate order while considering the order of selecting the target pixel from the plurality of pixels included in the raster group. Keep it. In this way, it is possible to read out the corresponding threshold values one after another without performing complicated address calculation, and the gradation number conversion process can be performed quickly.

  Further, in the tone number conversion process of the second embodiment, as in the first embodiment described above, a plurality of rasters are grouped as a raster group, and the target pixels are selected in a predetermined order, whereby each raster is parallelized. And processing. For this reason, as in the first embodiment, the cache hit rate, the efficiency of pipeline processing, and the predictive hit rate are improved, so that the processing speed can be increased.

  However, the phenomenon of improving the cache hit rate, pipeline processing efficiency, and branch prediction hit rate tends to be similar to the tone values of adjacent pixels in the image data. It is due to the facts. Therefore, if the number of rasters constituting the raster group becomes too large, the gradation values of the pixels in the raster group cannot be approximated. Therefore, the cache hit rate, the efficiency of pipeline processing, and branch prediction The degree of improvement in the hit rate is reduced, and the degree to which the processing speed can be increased is reduced. However, if each threshold value of the threshold value matrix is stored in consideration of the order of selecting the target pixel as in the gradation number conversion process of the second embodiment, for example, the cache hit rate and the efficiency of the pipeline process are stored. Even when the accuracy of branch prediction cannot be improved, the gradation number conversion process can still be performed quickly.

  Furthermore, the gradation number conversion process of the second embodiment can effectively improve the processing speed even in the following cases. For example, when decompressing image data compressed in the JPEG format, the raster data is not decompressed one by one, but is expanded into gradation value data of each pixel by taking 8 pixels × 8 pixels as a group. Therefore, the gradation values developed in this way are not rearranged in the order of the raster, but in the order as they are, in other words, in the order in which these pieces of data are developed into a plurality of pixels of 8 rows and 8 columns, If the tone number conversion process is performed for the above, the processing speed can be increased. At this time, the gradation values are developed in the same order each time for a plurality of pixels of 8 rows and 8 columns. Accordingly, in the second gradation number conversion process, if the threshold values in the threshold matrix are stored in an appropriate order in consideration of the order in which the gradation values are sequentially developed in the plurality of pixels, the processing speed is increased. Can be further improved. Specifically, the threshold matrix is grouped into units of 8 rows, and each unit is further divided into blocks of 8 rows and 8 columns. Then, the threshold values in the block are stored in the order in which the compressed image data is expanded. In this way, the threshold value can be read out without performing complicated address calculation when performing the gradation number conversion processing on the data developed in each pixel in the same order, so that the processing can be performed more quickly. Is possible.

  In the above description, the case where the image data is compressed in the JPEG format has been described as an example. However, there are various other cases where the processing speed can be increased for the same reason. For example, since the printing resolution of the printer is higher than the resolution of the image data, interpolation calculation is performed to generate image data for n pixels in the vertical direction and m pixels in the horizontal direction from the image data of one pixel. In this case, the processing can be speeded up by processing a plurality of pixels of n rows and m columns as a group, and the threshold values in the threshold matrix can be stored in consideration of the order of processing these pixels. Thus, the processing speed can be further increased.

C-3. Variations:
As described above, since the size of the threshold matrix is smaller than the size of the image data, the threshold matrix is repeatedly used while moving the position with respect to the image data in order to determine the presence or absence of dot formation for all pixels. . In the various embodiments described above, the threshold value matrix is moved in the direction along the raster. When moving in this way, focusing on one raster, the threshold value of the same part of the threshold value matrix is repeatedly used. For this reason, a certain pattern occurs in dot formation in the same raster depending on image data. This is particularly an output device having a variation element for each raster, for example, a serial scanning inkjet printer that prints a specific raster using a specific combination of nozzles instead of printing the raster using an arbitrary nozzle. When there is a variation in the dot landing position of a specific nozzle, it may overlap with a certain pattern in the raster direction and deteriorate the output image quality. In order to avoid the occurrence of such a problem, when the threshold value matrix is moved with respect to the image data, the threshold value matrix has a slight inclination with respect to the raster direction instead of being moved so as to completely coincide with the raster direction. It is effective to move it. In this way, when focusing on one raster, each time the position of the threshold matrix is moved, threshold values stored in slightly different portions of the matrix are used. Therefore, it is possible to reliably avoid the occurrence of a certain pattern in the formation of dots.

  In the gradation number conversion process of the second embodiment described above, the threshold value matrix is divided into a plurality of units, and the threshold values included in each unit are stored in an appropriate order. By shifting this in the direction intersecting with the raster and setting this shift amount to an appropriate value, the tone number conversion process can be performed more rapidly. Below, the modification of such 2nd Example is demonstrated.

  FIG. 15A shows a state in which the tone number conversion process is performed by repeatedly using one matrix while moving the threshold value matrix relative to the image data. In the drawing, a small rectangle indicated by a thin broken line schematically shows a pixel, and a thick broken line schematically shows a threshold matrix. Here, it is assumed that the raster group is composed of two rasters, and the tone number conversion process is performed while the two rasters are arranged in parallel. In FIG. 15A, the hatching given to the raster 1 and the raster 2 represents that the gradation number conversion processing is performed by collecting these rasters as a raster group.

  In the tone number conversion process shown in FIG. 15A, when the threshold value matrix is moved along the raster, the threshold value matrix is moved while being shifted by two rasters in the direction intersecting the raster. In more general terms, when the number of rasters constituting a raster group is N, the threshold value matrix is moved while being shifted by the number of rasters corresponding to an integral multiple of N in the upward or downward direction of the raster. The arrows shown in FIG. 15 (a) schematically show how the threshold value matrix is moved while shifting by two rasters. This is particularly true with the number of rasters constituting the raster group. The case where the shift amount of the threshold value matrix (the number of rasters to which the threshold value matrix shifts) matches is illustrated. In this way, if the shift amount of the threshold matrix is set to an integral multiple of the number of rasters constituting the raster group, the gradation number conversion process can be performed quickly. Hereinafter, this reason will be described with reference to the example of FIG.

  As described above, in the gradation number conversion process of the second embodiment, the threshold values of the threshold value matrix are stored in an appropriate order for each unit in consideration of the order of selecting the target pixel. Here, as shown in FIG. 15 (b), it is divided into three units, unit A to unit C, and stored in consecutive addresses on the memory of the computer 100.

  When the gradation number conversion process of image data is started, first, the presence or absence of dot formation for each pixel is determined while reading the threshold value included in the unit A of the threshold value matrix. Since the threshold value of the unit A is stored in advance in an appropriate order in consideration of the order of determining the presence / absence of dot formation, the appropriate threshold value can be read sequentially by simply reading while incrementing the address value one by one.

  Even if all the threshold values in unit A are read out, unprocessed pixels remain in the raster being processed, so the threshold value matrix is moved relative to the image data, and the presence or absence of dot formation is determined for these pixels. Continue processing. Here, as shown in FIG. 15A, when the threshold value matrix is moved in the raster direction, it is moved to a position shifted by two rasters in the direction intersecting the raster. The shift amount of the threshold value matrix corresponding to two rasters corresponds to the fact that the raster group is composed of two rasters. If the threshold value matrix is shifted in this way, the unit corresponding to the raster group being processed shifts, so that the threshold value to be read for determining the dot formation of the subsequent pixels is read from the unit B. Here, the threshold values in the unit B are also stored in advance in an appropriate order in consideration of the order in which the presence / absence of dot formation is determined. Therefore, the appropriate threshold values are sequentially read out by incrementing the address value one by one. be able to.

  When all the threshold values in the unit B are read out in this way, the threshold value matrix is moved again with respect to the image data as shown in FIG. At this time, it is moved in a state where it is shifted by two rasters in the direction intersecting the raster. By shifting the threshold value matrix in this way, the threshold value that is read to determine the presence / absence of dot formation for the subsequent pixels is the threshold value included in the unit C. When all the threshold values in the unit C are read, the threshold value matrix is moved again, and this time, the determination of dot formation is continued while reading the threshold values in the unit A.

  In FIG. 16, as described above, when the threshold matrix is moved relative to the image data, the threshold matrix is moved while being shifted in the direction intersecting the raster by the same number as the number of rasters constituting the raster group. It is explanatory drawing which showed typically a mode that the unit from which a threshold value is read switches sequentially.

  When processing the first raster group, as shown in the uppermost part of FIG. 16, the threshold value is first read from unit A, then read from unit B, and then read from unit C. Thereafter, the threshold values are read one after another in the order of unit A, unit B, and unit C. Here, as described above, since the threshold value is stored in a continuous address in each unit, the unit B is stored in a state where the unit B is merged with the unit A, that is, following the last address of the unit A, Further, if the unit C is stored in the state of being merged with the unit B, all the threshold values stored in the threshold value matrix can be read out continuously by incrementing the address value one by one.

  When the processing of all the pixels of the raster group being processed is completed and the processing of a new raster group is started, this time, from the top address of unit B, as shown in the second row from the top in FIG. The threshold value is read one after another while incrementing the address value one by one. In the figure, the dashed arrows displayed from the top row to the second row schematically show that the processing of the first raster group has been completed and the processing of the next raster group has been started. It is. When the processing of the second raster group is completed, the threshold value may be read one after another while incrementing the address value one by one from the head address of unit C. When the processing of the third raster group is completed, the process returns to unit A, and the threshold values are read in order from the head address of unit A. Thereafter, the same processing is repeated until it is determined whether or not dots are formed for all the pixels of the image data.

  As described above, when the threshold value matrix is moved with respect to the image data, it is complicated if it is moved in the direction intersecting with the raster while being shifted by the same number of rasters constituting the raster group as described above. Since an appropriate threshold value can be read sequentially without performing address calculation, the gradation number conversion process can be performed quickly.

  In the above description, the threshold value matrix is shifted upward with respect to the image data by the same number of rasters as the raster group. However, the present invention is not limited to this. For example, the threshold value matrix is shifted downward with respect to the image data. Even in this case, or even when the number of rasters constituting the raster group is shifted by an integral multiple, the gradation number conversion process can be speeded up. Hereinafter, such a case will be described.

  FIG. 17 is an explanatory view exemplifying a case where the threshold matrix is shifted while being shifted by the number of rasters corresponding to twice the number of rasters constituting the raster group, that is, by four rasters. FIG. 17A schematically illustrates a state in which a pixel (target pixel) to be used to determine whether or not dots are formed is associated with a threshold value stored in the threshold value matrix while moving the threshold value matrix with respect to the image data. Is shown. FIG. 17B schematically shows a state in which the units for reading out the threshold values change one after another as the threshold value matrix is moved. As shown in the figure, when the first raster group is processed, the threshold value of unit A is read out, and when all the threshold values of unit A are read out, the threshold value of unit C is read out next time. When all the threshold values of the unit C are read out, the unit A is returned to the unit A and read out in order from the first threshold value. Thereafter, it is determined whether or not dots are formed while repeatedly reading out the threshold values in the order of unit A and unit C until the processing for all the pixels in the first raster group is completed.

  When all the first raster groups have been processed, it is determined whether or not dots are formed for the next raster group while reading the threshold values repeatedly in the order of unit B and unit A. For the third raster group, the presence / absence of dot formation is determined while repeatedly reading out threshold values in the order of unit C and unit B. Thereafter, until the processing for all the pixels of the image data is completed, the presence or absence of dot formation may be determined while reading out the threshold values one after another.

  After all, if the threshold value matrix is shifted as shown in FIG. 17, it is sufficient to read out the threshold values stored in consecutive address values at least within the unit in order, and complicated address calculation is not required. The logarithmic conversion process can be performed quickly. Further, the order of reading out each unit can be repeated with the same pattern while processing one raster group, so that complicated address calculation is unnecessary. In the above description, the case where the threshold matrix is shifted by the number corresponding to twice the number of rasters constituting the raster group when moving the threshold matrix has been described. However, generally, the shift amount of the threshold matrix is the number of rasters constituting the raster group. If the integer multiple is set, the same thing is true, so that the gradation number conversion process can be speeded up.

D. Third embodiment:
In the various embodiments described above, the tone number conversion process is performed using a so-called dither method, but there is a so-called error diffusion method as a typical method for performing tone number conversion. . In the third embodiment described below, the tone number conversion process is performed by combining the dither method and the error diffusion method.

D-1. Tone number conversion processing of the third embodiment:
First, a basic concept when the gradation number conversion process is performed by combining the dither method and the error diffusion method will be described with reference to FIG. The color image data is usually expressed by combining image data of each color. In the example shown in FIG. 18, the color image data includes cyan (C) image data, magenta (M) image data, yellow (Y) image data, and black (K) image. Expressed as a combination of data. When the dither method and the error diffusion method are used in combination, these methods are used separately for each color. Here, the dither method has an advantage that it can be processed more quickly than the error diffusion method, but on the other hand, the image quality may be inferior to the error diffusion method. Therefore, the tone number conversion process is performed using the dither method for image data with a color that has relatively little influence on the image quality, and the tone number conversion process is performed using an error diffusion method for image data with other colors. Shall.

  FIG. 18B is an explanatory diagram conceptually showing a state in which image data of each color of C, M, Y, and K is read from the target pixel. A small rectangle indicated by a thin wavy line in the drawing schematically represents a pixel, and the target pixel is displayed with hatching. Among the image data of C, M, Y, and K colors, the Y color is the image data that has the least influence on the image quality. Therefore, in the third embodiment described below, the tone number conversion process is performed on the Y color image data using the dither method, and the tone number conversion is performed on the other color image data using the error diffusion method. Process.

  FIG. 19 is a flowchart showing the flow of processing for determining the presence or absence of dot formation while properly using the dither method and the error diffusion method for each color of C, M, Y, and K in the tone number conversion processing of the third embodiment. It is. Such processing is executed by the printer driver using the function of the CPU 102 of the computer 100 during the image data conversion processing shown in FIG.

  When the tone number conversion process of the third embodiment is started, first, a process of forming a raster group by combining a predetermined number of adjacent rasters is performed (step S300). Next, one target pixel from which a dot formation is determined is selected from a plurality of pixels constituting the raster group (step S302). And the process which reads the image data of each color of C, M, Y, and K memorize | stored matched with the object pixel is performed (step S304). Here, the image data is described as being composed of a combination of C, M, Y, and K color data. However, for example, light cyan (LC) or light magenta (LM) is used as color image data. When image data of other colors is included, the image data of each color is also read at the same time. The image data for each color may be stored at successive addresses on the memory of the computer 100, but may be stored in a different area for each color image data.

  When the image data of each color is read in this way, it is determined whether or not the image data is Y (step S306). This determination can be made based on, for example, the order in which the image data is read for the target image, or by adding identification data to the image data in advance and making this determination. If the image data is Y image data (step S306: yes), the presence or absence of dot formation for the target pixel is determined by applying the dither method (step S308). If the image data is not Y image data (step S306). : No) The presence or absence of dot formation is determined by applying the error diffusion method (step S308). Details of these processes will be described later. If the image data of each color includes LC or LM image data, in addition to the Y image data, the dot and dithering is applied to these LC and LM image data. Judging.

  When the presence / absence of dot formation for the target pixel is determined, it is determined whether or not dot formation has been determined for all pixels included in the raster group being processed (step S312), and unprocessed pixels remain. If (step S312: no), the process returns to step S302 and repeats a series of processes. When the determination is completed for all the pixels in the raster group (step S312: yes), it is next determined whether or not the processing for all the rasters included in the image data has been completed (step S314). If there is any raster that has not yet been processed, the process returns to step S300 to generate a new raster group, and the series of processes described above are performed on the raster group. When the processing for all the rasters is completed in this manner, the gradation number conversion process of the third embodiment is exited, and the process returns to the image data conversion process shown in FIG.

  FIG. 20 is a flowchart showing the flow of processing for determining the presence or absence of dot formation for the target pixel by applying the dither method during the tone number conversion processing of the third embodiment. FIG. 21 is an explanatory diagram schematically showing a state in which the presence / absence of dot formation is determined while selecting a target pixel during the gradation number conversion process of the third embodiment. The numbers shown in square brackets in FIG. 21 indicate the order in which the target pixel is selected. In FIG. 21, two adjacent rasters are grouped together to form a raster group. Hereinafter, a description will be given according to the flowchart of FIG. 20 with reference to FIG.

  In the process of applying the dither method to determine the presence / absence of dot formation, a process of reading the threshold value th corresponding to the target pixel being processed from the threshold matrix is performed (step S330). Any of the methods described in the various embodiments described above can be applied to the process of reading the threshold th. Although detailed description is omitted here, the threshold value is read as follows, for example. As an example, when the pixel numbered as [1] in FIG. 21 is selected as the target pixel, when the threshold matrix is superimposed on the image data, the square at the position of the pixel is selected from the threshold matrix. It is only necessary to select and read the threshold value stored in the cell.

  Next, the magnitude of the gradation value is compared between the threshold th read in step S330 and the image data of the target pixel (here, Y image data). The image data of the target pixel is already read in step S304 during the gradation number conversion process shown in FIG. If the image data is larger (step S332: yes), it is determined that a dot is to be formed on the target pixel, and a value “1” indicating that a dot is to be formed in the variable Dr indicating the determination result of the presence or absence of dot formation. Is written (step S334). On the contrary, when the threshold th is larger, a process of writing a value “0” indicating that no dot is formed is performed in the variable Dr indicating the determination result (step S336). When the presence / absence of dot formation is determined as described above, the processing for determining the presence / absence of dot formation by the dither method is skipped, and the processing returns to the tone number conversion processing shown in FIG. In this way, the target pixel selected in the gradation number conversion process is processed. As a result, with respect to the Y image data, it is possible to determine the presence or absence of dot formation by applying a plurality of rasters in parallel while applying the dither method as shown in FIG.

  For image data other than Y, the presence or absence of dot formation is determined using an error diffusion method. Here, since the target pixels are selected in the order shown in FIG. 21, the determination of dot formation by the error diffusion method is also performed in the order shown in FIG. Hereinafter, processing for determining whether or not dots are formed for a target pixel by applying the error diffusion method will be described.

  FIG. 22 is a flowchart showing a flow of processing for determining the presence / absence of dot formation by applying the error diffusion method. Such processing is performed for each color of C, M, and K. In order to avoid complicated description, the following description will be made without specifying a color.

  When processing by the error diffusion method is started, first, processing for reading the diffusion error er diffused in the target pixel is performed (step S360). As is well known, in the error diffusion method, a dot is formed for a certain pixel or a dot is not formed, so that an error in gradation expression (grayscale error) generated in that pixel is detected by surrounding unprocessed pixels. It spreads in association with. Then, when determining the presence or absence of dot formation for a new pixel, the presence or absence of dot formation is determined so as to eliminate the error in consideration of the error diffused in association with the pixel from surrounding pixels. . In step S360, a process of reading an error (diffusion error er) diffused from the pixel for which dot formation has already been determined toward the target pixel is performed. The process of diffusing the gradation expression error will be described later.

  Next, the correction data Dc is calculated by adding the read diffusion error er and the image data Dt of the target pixel (step S362). Then, the magnitude relationship between the obtained correction data Dc and the predetermined threshold value ther is determined (step S364). The threshold value ther is set to an appropriate value in advance. When it is determined that the correction data Dc is larger than the threshold ther (step S364: yes), it is determined that a dot is to be formed in the target pixel, and a value “meaning that a dot is formed in the variable Dr indicating the determination result”. 1 "is written (step S366). On the contrary, when the correction data Dc is smaller than the threshold value ther, a process of writing a value “0” indicating that no dot is formed is performed in the variable Dr indicating the determination result (step S368).

  When it is determined whether or not dots are to be formed for the target pixel in this manner, the gradation expression error (gradation error) caused by the determination is calculated, and then the obtained gradation error is diffused to surrounding undetermined pixels. Processing is performed (step S370). The gradation error is calculated by subtracting a gradation value (this gradation value is referred to as a result value) expressed in the target pixel as a result of determining whether or not dots are formed from the gradation value of the target pixel. For example, if it is determined that no dot is formed in the target pixel, no dot is formed in that pixel, and the result value is the gradation value “0”. If it is determined that a dot is to be formed on the target pixel, the gradation value represented by the pixel when the dot is actually formed (usually, the gradation value “255”). Therefore, the gradation error can be obtained by subtracting these result values from the gradation value of the target pixel.

  The gradation error calculated in this way is distributed to surrounding pixels with a predetermined weight. FIG. 23 is an explanatory diagram illustrating weighting factors used when distributing errors. Small rectangles in the figure indicate pixels. Further, a pixel marked with * indicates a target pixel, and a numerical value shown in each rectangle indicates a weighting coefficient. For example, in the example shown in FIG. 23A, the gradation error is distributed to the pixel on the right side of the target pixel using the weight coefficient ¼. In other words, a gradation value that is ¼ of the gradation error that has occurred in the target pixel is distributed to the pixel on the right. Similarly, a gradation value of ¼ of the gradation error is distributed to each pixel at the lower right, right below, and lower left of the target pixel. In addition, since it is already determined whether or not dots are formed on the upper or left adjacent pixel of the target pixel, no error is distributed to these pixels. As described above, if the gray scale error is diffused using the weighting coefficient shown in FIG. 23A, the gray scale error is distributed to each pixel. As a result, the four undecided pixels around the target pixel are divided into the target pixels. The gradation error generated in the pixels is distributed by 1/4. In this specification, “distributing gradation error” means adding a gradation value obtained by multiplying a gradation error by a predetermined weighting factor to a memory associated with a certain pixel. It shall be. “Diffusion of gradation error” means that gradation error is distributed to a plurality of pixels. Further, an error accumulated as a result of the distribution of gradation errors from surrounding pixels is referred to as a “diffusion error”.

  The weighting factors used for diffusing the gradation error are not limited to those illustrated in FIG. 23A, and various weighting factors, for example, as shown in FIG. Can be used. When the gradation error generated in the target pixel is diffused to the undetermined pixels in the vicinity (step S370 in FIG. 22), the dot formation determination process by the error diffusion method shown in FIG. Return to the tone conversion processing in the example.

  As described above, in the gradation number conversion process of the third embodiment, when the gradation number conversion process is performed on the data of each color constituting the color image data, the difference in the influence on the image quality due to the difference in color. In consideration of the above, gradation number conversion processing is performed while appropriately using the dither method and the error diffusion method. When the tone number conversion process is performed by applying the dither method, a plurality of adjacent rasters are collected as a raster group, and the tone number conversion process is performed on the rasters in parallel, thereby speeding up the processing. Can be achieved.

  In addition, in the tone number conversion process of the third embodiment, the determination by the error diffusion method is also performed in parallel with the determination by the error diffusion method, corresponding to the determination by the dither method being performed in parallel. . As a result, the processing by the error diffusion method can also be speeded up, so that the overall gradation number conversion processing can be further speeded up together with the speeding up of the processing by the dither method. Hereinafter, the reason why it is possible to quickly determine whether or not dots are formed by the error diffusion method by processing a plurality of rasters in parallel will be described.

  Here, a description will be given on the assumption that a raster group is composed of two rasters, and the target pixel is selected from the pixels included in the raster group in the order shown in FIG. FIG. 24 is an explanatory diagram showing a state in which target pixels are sequentially selected from the pixels included in the raster 3 and the raster 4 to determine whether or not dots are formed. A small rectangle in the figure indicates a pixel, and the numbers described in square brackets in the pixel indicate the order in which the pixel is processed in the raster group.

  FIG. 24A schematically shows a state in which gradation errors are diffused from the target pixel first selected in the raster group (the pixel numbered [1] in the figure) to the surrounding pixels. ing. A solid line arrow in the drawing schematically represents a state in which gradation errors are distributed to surrounding pixels according to a predetermined weighting factor. Here, it is assumed that the error is diffused according to the weighting factor shown in FIG. When the gradation error generated in the pixel with the number [1] is diffused, a process for determining whether or not dots are formed is started for the pixel with the number [2] on the right. That is, the diffusion error stored in association with the pixel with the number [2] is read to calculate correction data, and compared with a predetermined threshold value to determine the presence / absence of dot formation. Then, when the gradation error caused by the determination is calculated, the obtained gradation error is diffused to surrounding undetermined pixels. FIG. 24B schematically shows a state in which the gradation error caused by the number [2] is diffused to surrounding pixels. In FIG. 24B, pixels that have already been determined for dot formation are hatched and displayed.

  Next, it is determined whether or not dots are formed for the pixel numbered [3]. In this determination, the diffusion error distributed to the pixel [3] from the surrounding pixels is read, and correction data for the code [3] is calculated to determine the presence / absence of dot formation. Here, as is clear from FIG. 24C, the gradation error distributed to the pixel of the number [3] is the error from the pixel of the number [1] and the pixel from the number [2]. It is only an error. The error distributed from the pixel with the number [2] has just been calculated in the immediately preceding process, and the error distributed from the pixel with the number [1] has just been calculated in the previous process. is there. Accordingly, all of these errors remain in the cache memory of the computer 100, so that they can be read out at a very high speed. If the error can be read out at high speed in this way, the correction data can be quickly calculated and the dot formation presence / absence determination for the number [3] can be quickly made. Here, the error value distributed from the number [1] or the number [2] has been described as being read from the cache memory. However, instead of reading the distributed error, the error occurred in these pixels. The gradation error may be read from the cache memory and the error distributed to the number [3] may be calculated. These processes can be considered as substantially equivalent processes.

  In the above description, the case where the dot formation presence / absence is determined for the pixel with the number [3] has been described. However, when the determination is made for the pixel with the number [2], the processing speed is increased for the same reason. It is possible. That is, when determining the presence or absence of dot formation for the pixel with the number [2], the diffusion error accumulated in association with the pixel with the number [2] is read from the memory to calculate correction data. In the diffusion error accumulated in the pixel of the number [2], errors from a plurality of pixels including the pixel of the number [1] are also accumulated, but the error from the pixel of the number [1] It is calculated by processing. Therefore, this error may be temporarily stored in the memory instead of being added to the diffusion error already stored in the memory in association with the number [2]. Then, a diffusion error for which the error from the number [1] is not yet distributed is read from the memory, and the error stored temporarily and added on the CPU 102, thereby the diffusion error for the number [2]. May be calculated. In this way, the error distributed from the pixel with the number [1] to the pixel with the number [2] is added to the diffusion error already accumulated in the memory associated with the pixel with the number [2]. Since processing is not necessary, it is possible to quickly perform processing for determining the presence or absence of dot formation.

  Note that the error stored in the memory independently without being added to the error from other pixels in this way is temporarily stored, and is used for the determination of dot formation. Therefore, since many of these errors are not stored, these errors can be stored in a memory such as a register that can be read and written at high speed. If the data is stored in a readable / writable memory in this manner, the process for determining the presence or absence of dot formation can be performed more quickly.

  In FIG. 24A, each pixel that distributes an error from the pixel with the number [2] is a pixel that determines whether or not a dot is formed. Therefore, these errors may be temporarily stored in the memory as errors, instead of being stored in the memory associated with the peripheral pixels. In this way, for the same reason as described above, it is possible to perform the process of determining the dot formation more quickly.

  On the other hand, as shown in FIG. 24C, the gradation error generated in the pixel with the number [3] is diffused into the raster 4 pixel and the raster 5 pixel. Here, since two rasters, that is, raster 3 and raster 4 are processed in parallel, the presence or absence of dot formation is not determined for a while for the pixels included in raster 5. . Therefore, the error distributed from the pixel with the number [3] to the pixels of the raster 5 is distributed to the memories associated with these pixels. When errors from other pixels have been distributed to these pixels, a value added to the error that has already been stored may be stored in the memory. The solid arrow in FIG. 24 indicates that the error value is temporarily stored in the memory instead of being distributed in association with the pixel, and the white arrow indicates that the error is transferred to the surrounding undetermined pixels. It shows that the data is distributed on the memory in association with each other.

  When the process for the pixel with the number [3] is completed, the process for the pixel with the number [4] is started. In the processing, the image data and the diffusion error stored in association with the number [4] are read from the memory. Then, correction data is calculated by adding an error from the pixel on the left (the pixel with the number [2]), and whether or not dots are formed is determined based on the magnitude relationship between the obtained correction data and a predetermined threshold value. to decide. Next, an error to be distributed to each pixel is calculated based on the gradation error generated by the determination and a predetermined weighting factor, and the obtained error is temporarily stored in the memory. In this way, when the process for the pixel of the number [4] is completed, the same process is repeated for the subsequent pixels of the number [5].

  As described above, in the tone number conversion process of the third embodiment, it is possible to increase the speed of the process using the dither method by processing a plurality of rasters constituting the raster group in parallel. Instead, it is possible to speed up the process of determining the presence or absence of dot formation by applying the error diffusion method.

D-2. Variations:
In the gradation number conversion process of the third embodiment described above, by processing a plurality of rasters constituting a raster group in parallel, both the process applying the dither method and the process applying the error diffusion method are performed. It is also possible to improve the processing speed. However, the method of processing a plurality of rasters included in the raster group in parallel is not limited to the method described above. In the modification described below, a plurality of rasters are processed in parallel by dividing the raster group into a plurality of blocks and determining whether or not dots are formed for each block. This also makes it possible to speed up the gradation number conversion process. FIG. 25 is a flowchart showing the tone number conversion process as a modification of the third embodiment. FIG. 26 is an explanatory view exemplifying a state in which the tone number conversion processing according to the modification converts image data into data in an expression format based on the presence or absence of dot formation. Hereinafter, a description will be given according to the flowchart shown in FIG. 25 with reference to FIG.

  When the tone number conversion process of the modified example of the third embodiment is started, first, a raster group is generated by collecting N adjacent rasters (step S400). FIG. 26 shows a state in which two rasters, raster 1 and raster 2, are combined into a raster group, assuming that the raster group is composed of two rasters.

  Next, a predetermined number of adjacent pixels in the raster group are grouped into a plurality of blocks, and a processing block for determining the presence / absence of dot formation is set (step S402). The block is represented by a rectangle indicated by a thick broken line in FIG. In the example shown in the figure, the raster group is divided into a plurality of blocks each consisting of a total of four pixels of two pixels vertically and horizontally. In step S402, the leftmost block among the plurality of blocks thus divided is set as a block for determining the presence or absence of dot formation (that is, a processing block).

  Image data of each color of C, M, Y, and K is read from each pixel included in the processing block thus set (step S404). That is, in the example shown in FIG. 26, the pixel numbered [1], the pixel numbered [2], the pixel numbered [3], and the pixel numbered [4] in the figure. The image data of each color is read for a total of four pixels. Next, it is determined whether the image data is Y-color image data (step S406). If the image data is Y-color image data (step S406: yes), the dither method is applied to each pixel in the processing block. It is determined whether or not there is any dot formation (step S408). That is, the presence or absence of dot formation may be determined by applying the dither method to the four pixels included in the processing block in order from the youngest pixel numbered in FIG. As specific methods for determining the presence or absence of dot formation, the methods of the various embodiments described above can be applied. Details of the process are the same as those described above, and thus the description thereof is omitted here.

  In step S406, if the image data is not Y-color image data (step S406: no), the presence / absence of dot formation is determined for each block while applying the error diffusion method (step S410). Details of this processing will be described later.

  When the presence or absence of dot formation for the processing block is determined in this way, it is determined whether or not the processing has been completed for all pixels included in the raster group being processed (step S412). If it is determined that an unprocessed pixel remains in the raster group, the process returns to step S402, and a series of processes are repeated until the process for all pixels is completed. If it is determined that all the pixels of the raster group have been processed (step S412: yes), it is determined whether or not the processing for all the rasters included in the image data has been completed (step S414). ). If an unprocessed raster remains (step S414: no), the process returns to step S400 to perform a series of subsequent processes. When it is determined that the processing has been completed for all the rasters included in the image data (step S414: yes), the gradation number conversion processing of the modification shown in FIG. 25 is ended, and the image data conversion processing of FIG. Return to.

  Hereinafter, the process of step S410, that is, the process of determining the presence / absence of dot formation for each pixel in the processing block while applying the error diffusion method will be described. The processing described below is performed in exactly the same manner for the C, M, and K color image data.

  FIG. 27 is a flowchart showing a flow of processing for determining whether or not dots are formed for each pixel in the processing block while applying the error diffusion method. When the processing is started, first, a sum S of gradation values for each pixel included in the processing block is calculated (step S450). The gradation value for each pixel in the processing block has already been read in the gradation number conversion process of the modification of the third embodiment described above with reference to FIG.

  Next, processing for reading the diffusion error er stored in association with each pixel in the processing block is performed (step S452). For example, in the case of processing the first block among the blocks shown in FIG. 26, the diffusion error er distributed toward each pixel of the number [1] to the number [4] is read out. do.

  The correction data Bx for the processing block is calculated by adding the diffusion error er of each pixel read out in this way and the previously calculated sum S (step S454). Then, the obtained correction data Bx is compared with a predetermined threshold th1 (step S456). If the correction data Bx is smaller than the threshold value th1, it is determined that no dot is formed in any pixel of the processing block (step S458).

  On the other hand, when it is determined in step S456 that the correction data Bx is larger than the threshold value th1, the comparison with the predetermined threshold value th2 is further performed (step S460). Here, the threshold value th2 and the threshold value th1 are set so that the relationship of th1 <th2 is satisfied. When the value of the correction data Bx is smaller than the threshold th2, that is, larger than the threshold th1 but smaller than the threshold th2 (step S460: no), only one pixel among the pixels included in the processing block is It is determined that dots are formed (step S462).

  FIG. 28 shows a state where dots are formed only in one of the four pixels included in the processing block. There are four methods shown in FIG. 28 for forming dots in only one pixel in the processing block. In this embodiment, in the process of step S462, when dots are formed only for one pixel in the processing block, dots are formed at pixels at positions selected randomly from these four patterns each time. Needless to say, dots may be formed at the same position each time, or dots may be formed at the pixel having the largest gradation value among the pixels included in the processing block.

  As described above, if it is determined that dots are not formed on the pixels included in the processing block (step S458), or if it is determined that dots are formed only on one pixel (step S462), such determination causes the processing block. An error in the gradation expression is calculated (step S464). The gradation error generated in the processing block can be calculated by subtracting the result value for the block from the gradation value of the correction data Bx of the processing block. Here, the result value for the processing block is the sum of the result values (tone values expressed in the pixels when dots are formed or not formed) for each pixel constituting the block. Value.

  For example, when no dot is formed in any pixel in the processing block (in the case of step S458), since the result value of each pixel is “0”, the result value of the target block is also “0”. Become. Accordingly, in the processing block, the gradation value of the correction data Bx is directly generated as a gradation error. Similarly, when a dot is formed only on one pixel in the processing block (in the case of step S462), the result value of the processing block is the result value for the pixel on which the dot is formed. Therefore, if the result value of the pixel in which the dot is formed is subtracted from the correction data Bx, the gradation error in the processing block can be calculated.

  When the gradation error generated in the processing block is calculated in this way, processing for diffusing this error to surrounding pixels is performed (step S466). FIG. 29 is an explanatory diagram conceptually showing a state in which a gradation error generated in a processing block is diffused to surrounding undetermined pixels. The large rectangles shown with hatching in the figure indicate processing blocks in which gradation errors have occurred, and the small rectangles indicated by solid lines around the processing blocks indicate peripheral pixels that distribute gradation errors. Yes. For example, when the gradation error is diffused to pixels adjacent to the processing block, the error is distributed to each pixel using the weighting factor described in each pixel in FIG. In addition, a weighting factor as illustrated in FIG. 29B may be used so that an error is evenly diffused to pixels around the processing block.

  Note that a large rectangle indicated by a thin broken line in FIG. 29A indicates a peripheral block adjacent to the processing block. In the example shown in FIG. 29A, the weighting coefficient of each pixel is determined so that the gradation error generated in the processing block is evenly distributed to adjacent peripheral blocks. If the weighting factor is set in this way, errors can be evenly distributed to the respective blocks, so that relatively good image quality can be easily obtained. Of course, when an error is distributed to a part of pixels in an adjacent block, it is expected that a dot is formed differently in a pixel in which the error is distributed compared to a pixel in which the error is not distributed. However, as apparent from the flowchart shown in FIG. 27, the gradation errors generated in the processing block are diffused together only in a region where the gradation value of the image data is small, in other words, the dots are sparse. This is a region that is not formed. In such a region, even if the dot formation is slightly different in each pixel in the block, the image quality is not greatly affected.

  It should be noted that the weighting factors shown in FIG. 29 are exemplary, and therefore, it is possible to obtain better image quality by correcting the weighting factors for each pixel while confirming the obtained image. Needless to say.

  If it is determined in step S460 of FIG. 27 that the correction data Bx of the processing block is larger than the predetermined threshold th2 (step S460: yes), a dot is applied to each pixel in the processing block while applying the error diffusion method. Processing for determining the presence or absence of formation is performed (step S468). The contents of such processing will be described with reference to FIG.

  FIG. 30 is an explanatory diagram schematically showing a state in which the presence / absence of dot formation is determined for each pixel in the processing block while applying the error diffusion method to diffuse the gradation error to surrounding pixels. By performing such processing for each block indicated by a thick dashed rectangle in FIG. 26, it is possible to determine the presence or absence of dot formation for each pixel. In order to generalize the description, in the following, the upper left pixel in the processing block is referred to as pixel Pa, the upper right pixel is referred to as pixel Pb, the lower left pixel is referred to as pixel Pc, and the lower right pixel is referred to as pixel Pd. Therefore, the pixel Pa corresponds to a pixel to which reference numerals [1], [5], [9],. Similarly, the pixel Pb corresponds to the pixels with the numbers [2], [6], [10]... In FIG. 26, and the pixels Pc are the numbers [3], [7], [11. ..., And the pixel Pd corresponds to the pixels with the numbers [4], [8], [12].

  In FIG. 30, a large rectangle indicated by a broken line indicates a processing block, and a small rectangle indicated by a solid line in the processing block indicates each pixel in the block. FIG. 30A schematically shows a state in which the gradation error ERa generated by determining the presence or absence of dot formation for the pixel Pa is diffused to the surrounding undetermined pixels. Here, since processing is performed for each block, the pixel on the left side of the processing block has already been determined for dot formation. In FIG. 30, the pixels for which dot formation has been determined are hatched and displayed. Therefore, as shown in FIG. 30A, the gradation error ERa generated in the pixel Pa is diffused to the three pixels in the processing block with a predetermined weight coefficient. FIG. 31 is an explanatory diagram exemplifying weighting factors for converting gradation errors. The * mark in the figure indicates a pixel in which a tone error has occurred due to the determination of dot formation, and the numerical values in the surrounding pixels indicate weighting factors for distributing the error to that pixel. ing.

  By multiplying the gradation error generated in the pixel Pa by the weighting coefficient shown in FIG. 31A, an error to be distributed to each pixel in the processing block can be calculated. Here, since each pixel to which an error is to be distributed is a pixel for which it is determined whether or not a dot is formed, these errors are temporarily stored in a memory. In FIG. 30, a solid line arrow pointing from a pixel to an adjacent pixel indicates that the calculated error is not temporarily distributed to the surrounding pixels and stored in the memory, but the error itself is temporarily stored. ing.

  Following the pixel Pa, the presence / absence of dot formation of the pixel Pb is determined. In the determination, the image data of the pixel Pb, the diffusion error stored in association with the pixel Pb, and the error temporarily stored in the memory (a predetermined weighting factor is applied to the gradation error ERa of the pixel Pa). (Multiplied gradation value) is read out and added to calculate correction data. By comparing the correction data thus obtained with a predetermined threshold value, it is possible to determine whether or not dots are formed for the pixel Pb. Next, the gradation error ERb generated in the pixel Pb is calculated according to the dot formation determination result, and the error to be distributed to the surrounding undetermined pixels is calculated using the weighting factor shown in FIG. The pixel to which the error thus obtained is to be distributed is considered to be a pixel for determining whether or not to form dots soon. Therefore, all the obtained errors are temporarily stored in the memory. The weighting factor shown in FIG. 31 (b) is such that more errors are distributed to the pixels in the block than the pixels outside the block being processed in consideration of performing the error diffusion method in units of blocks. Is set to Of course, a weighting coefficient that evenly distributes errors to surrounding pixels may be used.

  Here, the reason why the error distributed from the pixel Pb to the right adjacent pixel and the lower right pixel is temporarily stored in the memory will be supplementarily described. As described with reference to FIG. 27, whether the determination of dot formation on a processing block is performed for each pixel or whether the blocks are collectively performed as described with reference to FIG. It is determined by the gradation value of the data. Therefore, regarding the block on the right side of the processing block, the presence / absence of dot formation is not necessarily determined for each pixel. That is, the error distributed from the pixel Pb to the right adjacent pixel or the lower right pixel is used soon when the dot formation is determined for each pixel in the right adjacent block. However, in most cases, the image data takes almost the same gradation value in the neighboring pixels. Therefore, if the block being processed is judged for each pixel, the pixel in the right adjacent block has a considerably high probability. It can be considered to judge every time. If so, the error to be distributed from the pixel Pb to the right adjacent pixel and the lower right pixel can also be temporarily stored in the memory, so that the process of determining whether or not dots are formed can be sped up. It is. In consideration of this, here, the error to be distributed from the pixel Pb to the peripheral pixels is temporarily stored in the memory. If it is determined whether the right adjacent block collects the blocks and whether or not dots are formed, the error temporarily stored in the memory may be distributed to these pixels. Further, more simply, the processing may be performed without distributing the temporarily accumulated error to these pixels. In the area where the presence / absence of dot formation as a whole block is determined, dots are formed sparsely, so even if error diffusion from adjacent blocks is omitted, the effect of this is limited and the processing is simplified. By doing so, the processing speed can be further improved.

  When the process for the pixel Pb is completed as described above, the process for the pixel Pc is started. Similarly, the presence or absence of dot formation is also determined for the pixel Pc, and the gradation error ERc caused by the determination is calculated. Next, the error to be distributed to the surrounding pixels is calculated according to the weighting factor shown in FIG. When the error from the pixel Pc is diffused, only the pixel Pd on the right side exists in the block being processed. Therefore, for the diffusion of the gradation error of the pixel Pc, the weighting coefficient shown in FIG. 31C is used so that the error is evenly distributed to the surrounding pixels. Of course, a weighting factor set so that more errors are distributed to the pixels in the block being processed may be used.

  Regarding the pixel Pd adjacent to the right among the peripheral pixels of the pixel Pc, the presence / absence of dot formation is determined immediately thereafter. On the other hand, the determination of dot formation for other pixels, that is, the lower left pixel, the lower right pixel, and the lower right pixel is after the determination for all the pixels included in the raster group being processed is completed. . Therefore, the error to be distributed from the pixel Pc to the right adjacent pixel is temporarily stored in the memory, and the other errors are distributed to the respective pixels. In other words, a predetermined address on the memory is stored in association with each pixel, and when errors from other pixels are distributed, the errors are accumulated in addition to the errors already stored. . As described above, in FIG. 30, the solid line arrow from the pixel to the adjacent pixel indicates that the calculated error is temporarily stored. On the other hand, from the pixel to the adjacent pixel, The white arrow heading indicates that the error is not temporarily stored but distributed to these pixels.

  When the process for the pixel Pc is completed in this way, it is determined whether or not dots are formed for the pixel Pd. In the example shown in FIG. 30, the error distributed to the pixel Pd is only the error from the three pixels Pa to Pc, and these errors are temporarily stored in the memory as described above. Yes. Therefore, when determining the pixel Pd, the correction data can be calculated by reading the image data of the pixel Pd and adding the error temporarily stored in the memory. This eliminates the need to read out the diffusion error stored in association with the pixel Pd from the memory, so that the correction data can be calculated quickly, and the determination of dot formation presence / absence can be quickly performed. Can be done. Further, the error to be temporarily stored has a limited number of data, and therefore can be stored in a memory such as a register that can be read and written at high speed. In this way, it is possible to further increase the speed. Furthermore, even if these errors are not intentionally stored in registers but stored in general memory, in practice, these errors are likely to be read from the cache memory, and are therefore usually read quickly. be able to. For this reason, it is possible to calculate correction data quickly, and it is possible to speed up the process of determining whether or not dots are formed.

  When it is determined whether or not dots are formed for the pixel Pd, the gradation error ERd generated by the determination is calculated and distributed to surrounding pixels. FIG. 30D schematically shows a state where the gradation error ERd generated in the pixel Pd is diffused. The error to be distributed to each pixel is calculated using the weighting factor shown in FIG. Then, the calculated errors are distributed to the lower left, right below, and lower right pixels of the pixel Pd, and the error to be distributed to the right adjacent pixel of the pixel Pd is temporarily stored in the memory.

  As shown in FIG. 30D, there is an undetermined pixel in the upper right of the pixel Pd. Therefore, the gradation error generated in the pixel Pd may be distributed to the upper right pixel. As described above, if the error is diffused also in the upper right pixel, the image quality may be effectively improved. FIG. 31D illustrates the setting of weighting factors used in such a case. Since the upper right pixel of the pixel Pd is a pixel that determines whether or not dots are formed following the pixel Pd, the error to this pixel is temporarily stored in the memory. In FIG. 30D, a broken-line arrow from the pixel Pd to the upper right pixel indicates that an error for this pixel is temporarily stored in the memory.

  In step S468 shown in FIG. 27, processing for determining whether or not dots are formed for each pixel in the processing block is performed as described above while applying the error diffusion method. In this manner, when it is determined whether or not dots are formed for all the pixels in the processing block, the process shown in the flowchart of FIG. 27 is skipped, and the gradation number conversion process of the modified example of the third embodiment shown in FIG. 25 is performed. Return. In the tone number conversion process, it is determined whether or not the process has been completed for all the pixels of the raster group being processed (step S412 in FIG. 25). If there are any unprocessed pixels, one process block is determined. Only the right direction is moved, and the following series of processes is repeated.

  As described above, in the tone number conversion process of the modification of the third embodiment, the raster group is divided into a plurality of blocks, and the presence / absence of dot formation is determined for each block. In this way, in addition to processing a plurality of rasters in the raster group in parallel, the tone number conversion processing is speeded up by determining whether or not dots are formed for each pixel in the block at once. be able to. For example, if it is determined whether or not dots are formed without distinguishing the pixels in the block according to the correction data of the block to be processed, the processing can be speeded up.

  In the above description, the correction data Bx for the processing block is calculated, and the obtained correction data Bx is compared with a predetermined threshold value, so that the processing method for the block is selectively used. May be determined based on the sum S in the block. In this way, since it is possible to determine the presence or absence of dot formation without calculating correction data, the processing can be speeded up.

  In the above description, the block is composed of 4 pixels in total of 2 pixels in the vertical and horizontal directions. However, the configuration of the block may be various as required. For example, it may be configured by more pixels, and the number of pixels may be different in the vertical direction and the horizontal direction.

  Furthermore, the order in which each pixel is processed in the block is not limited to the order described above, and can be set to an appropriate order as necessary.

  Further, the range in which the gradation error is diffused is not limited to that illustrated in FIG. 29 or FIG. 31, and may be diffused to a farther pixel.

  Although various embodiments have been described above, the present invention is not limited to all the embodiments described above, and can be implemented in various modes without departing from the scope of the invention. For example, a software program (application program) that realizes the above functions may be supplied to a main memory or an external storage device of a computer system via a communication line and executed. Of course, a software program stored in a CD-ROM or a flexible disk may be read and executed.

  In the various embodiments described above, the image data conversion process including the gradation number conversion process is described as being executed in the computer. However, part or all of the image data conversion process is performed on the printer side or a dedicated image. It may be executed using a processing device.

  Furthermore, the image display device is not necessarily limited to a printing device that prints an image by forming ink dots on a print medium. For example, the present invention can be effectively applied to a liquid crystal display device that expresses an image whose gradation changes continuously by dispersing bright spots at an appropriate density on the liquid crystal display screen. It is.

1 is an explanatory diagram showing an outline of an image processing apparatus according to the present invention, taking a printing system as an example. FIG. It is explanatory drawing which shows the structure of the computer as an image processing apparatus of a present Example. 1 is a schematic configuration diagram of a printer used in combination with the image processing apparatus of the present embodiment. It is explanatory drawing which shows a mode that the several nozzle is provided in the bottom face of the printing head of a printer. It is a flowchart which shows the flow of the image data conversion process performed with the image processing apparatus of a present Example. It is explanatory drawing which illustrated the threshold value matrix used in an image data conversion process. It is explanatory drawing which shows a mode that the threshold value memorize | stored in the threshold value matrix is matched with image data, and the presence or absence of dot formation is judged. 6 is a flowchart showing a flow of processing for determining the presence or absence of dot formation while paralleling a plurality of rasters in the gradation number conversion processing of the first embodiment. It is explanatory drawing which showed typically a mode that the presence or absence of dot formation was judged in parallel with several rasters in the gradation number conversion process of 1st Example. It is explanatory drawing which shows a mode that the presence or absence of dot formation is judged in the various modifications of the gradation number conversion process of 1st Example, making several rasters parallel. It is explanatory drawing which shows the display method of each square which comprises a threshold value matrix. It is explanatory drawing which showed notionally the mode that the threshold value of a threshold value matrix was stored in consideration of the order which selects a symmetrical pixel in the gradation number conversion process of 2nd Example. It is explanatory drawing which showed typically a mode that the threshold value was read from the threshold value matrix in the gradation number conversion process of 2nd Example. It is explanatory drawing which shows a mode that the threshold value of a threshold value matrix is memorize | stored in the suitable order according to the order which selects an object pixel. FIG. 10 is an explanatory diagram showing a state in which a threshold value matrix is associated with image data while being shifted by the same number of rasters as the number of rasters included in a raster group in a modification of the second embodiment. It is explanatory drawing which shows a mode that a threshold value is read from a threshold value matrix, when it matches, shifting a threshold value matrix by the same raster number as the number of rasters contained in a raster group. FIG. 15 is an explanatory diagram showing a state in which a threshold value matrix is associated with image data while being shifted by a number of rasters that is an integral multiple of the number of rasters included in a raster group in another modification of the second embodiment. It is explanatory drawing which shows the basic way of thinking at the time of performing the gradation number conversion process combining the dither method and the error diffusion method. It is a flowchart which shows the flow of the gradation number conversion process of 3rd Example which processes while switching a dither method and an error diffusion method. It is a flowchart which shows the flow of the process which judges the dot formation presence or absence by applying a dither method in the gradation number conversion process of 3rd Example. It is explanatory drawing which showed typically a mode that the presence or absence of dot formation was judged in the gradation number conversion process of 3rd Example, making several rasters parallel. It is a flowchart which shows the flow of the process which judges the dot formation presence or absence by applying an error diffusion method in the gradation number conversion process of 3rd Example. It is explanatory drawing which illustrated the weighting coefficient used in order to diffuse a gradation error to the surrounding undetermined pixel in an error diffusion method. It is explanatory drawing which shows typically a mode that the presence or absence of dot formation is judged by applying an error diffusion method, making the several raster parallel in the gradation number conversion process of 3rd Example. It is the flowchart which showed the flow of the gradation number conversion process of the modification of the 3rd Example which divides | segments a raster group into a some block, and processes. It is explanatory drawing which showed typically a mode that the raster group was divided into the some block in the modification of 3rd Example. It is a flowchart which shows the flow of the process which judges the dot formation presence or absence for every block using an error diffusion method in the gradation number conversion process of the modification of 3rd Example. FIG. 10 is an explanatory diagram illustrating a state in which whether or not dots are formed is determined collectively in blocks in a gradation number conversion process according to a modified example of the third embodiment. It is explanatory drawing which illustrated the weighting coefficient for diffusing the gradation error which generate | occur | produced in the block to a surrounding pixel in the gradation number conversion process of the modification of 3rd Example. FIG. 10 is an explanatory diagram showing a state in which the presence / absence of dot formation is determined for each block while applying an error diffusion method in a tone number conversion process of a modification of the third embodiment. It is explanatory drawing which illustrated the weighting coefficient for diffusing the gradation error produced in each pixel in a block to a surrounding pixel in the gradation number conversion process of the modification of 3rd Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Image processing apparatus 20 ... Printer 100 ... Computer 102 ... CPU
104 ... ROM
106 ... RAM
108 ... Peripheral device interface PI / F
109 ... Disk controller DDC
110: Network interface card NIC
112 ... Video interface V-I / F
114 ... CRT
116: Bus 118 ... Hard disk 120 ... Digital camera 122 ... Color scanner 124 ... Flexible disk 126 ... Compact disk 200 ... Color printer 230 ... Carriage motor 235 ... Paper feed motor 236 ... Platen 240 ... Carriage 241 ... Print heads 242, 243 ... Ink Cartridge 244 ... Ink ejection head 260 ... Control circuit 261 ... CPU
262 ... ROM
263 ... RAM
300 ... Communication line 310 ... Storage device

Claims (11)

An image processing apparatus that converts the color image data into data of each color in an expression format based on the presence or absence of dot formation by determining the presence or absence of dot formation for each pixel constituting each color image data,
Raster group generation means for generating a raster group by combining predetermined N (N is an integer of 2 or more) rasters that are adjacent to each other, which are rasters of pixel rows;
For the image data of the first color component selected in advance from the color components, the raster group is configured by comparing the gradation value of each pixel with a predetermined threshold value associated with each pixel. First data conversion means for performing a first image conversion process of converting the N rasters to dot rows in parallel, thereby converting the data into data in an expression format depending on whether or not dots are formed;
For the image data of the second color component obtained by removing the first color component from each color component, the gradation error caused by the pixel determined by determining the presence / absence of dot formation is diffused to surrounding undetermined pixels. At the same time, the N rasters constituting the raster group are converted into dot rows in parallel while judging the presence / absence of dot formation so as to eliminate the gradation error diffused from the surrounding judged pixels. And a second data conversion means for performing a second image conversion process for converting the data into an expression format data based on the presence or absence of dot formation,
The first and second data conversion means perform the first and second image conversion processing as processing of a computer that implements at least one of cache and pipeline processing ,
The first data conversion means includes
Threshold matrix storage means for storing a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged;
Target pixel selection means for selecting a target pixel for determining whether or not to form a dot;
A dot formation determination unit that reads out a threshold value corresponding to the target pixel from the threshold value matrix and compares the threshold value with the gradation value corresponding to the target pixel, thereby determining whether or not the target pixel is formed. When
With
The target pixel selecting means is means for selecting the target pixel in parallel from each raster constituting the raster group by sequentially selecting each pixel included in the raster group in a predetermined order.
An image processing apparatus in which the pixel selection order in the target pixel selection means is up to two pixels that are continuously selected on the same raster . An image processing apparatus that converts the color image data into data of each color in an expression format based on the presence or absence of dot formation by determining the presence or absence of dot formation for each pixel constituting each color image data,
Raster group generation means for generating a raster group by combining predetermined N (N is an integer of 2 or more) rasters that are adjacent to each other, which are rasters of pixel rows;
For the image data of the first color component selected in advance from the color components, the raster group is configured by comparing the gradation value of each pixel with a predetermined threshold value associated with each pixel. First data conversion means for performing a first image conversion process of converting the N rasters to dot rows in parallel, thereby converting the data into data in an expression format depending on whether or not dots are formed;
For the image data of the second color component obtained by removing the first color component from each color component, the gradation error caused by the pixel determined by determining the presence / absence of dot formation is diffused to surrounding undetermined pixels. At the same time, the N rasters constituting the raster group are converted into dot rows in parallel while judging the presence / absence of dot formation so as to eliminate the gradation error diffused from the surrounding judged pixels. A second data conversion means for performing a second image conversion process for converting the data into an expression format data depending on whether or not dots are formed.
With
The first and second data conversion means perform the first and second image conversion processing as processing of a computer that implements at least one of cache and pipeline processing,
The first data conversion means includes
Threshold matrix storage means for storing a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged;
Target pixel selection means for selecting a target pixel for determining whether or not to form a dot;
A dot formation determination unit that reads out a threshold value corresponding to the target pixel from the threshold value matrix and compares the threshold value with the gradation value corresponding to the target pixel, thereby determining whether or not the target pixel is formed. When
With
The target pixel selecting means is means for selecting the target pixel in parallel from each raster constituting the raster group by sequentially selecting each pixel included in the raster group in a predetermined order.
An image processing apparatus in which the selection order of pixels in the target pixel selection means shifts to selection of pixels of adjacent rasters after sequentially selecting pixels at the same position in the raster direction in the plurality of rasters. The image processing apparatus according to claim 1 or 2 ,
The second data conversion means includes
The last raster in the raster group is selected, and a dot string representing the presence or absence of dot formation is determined by determining the presence or absence of dot formation for each pixel constituting the last raster. The last raster converting means for converting to
A first error diffusing means for calculating the gradation error for each pixel constituting the tail raster and diffusing to a plurality of undetermined pixels around each pixel;
The first raster at the head position in the raster group adjacent to the last raster is selected, and the first raster is selected while considering the gradation error diffused from the last raster to each pixel of the first raster. First raster conversion means for converting the first raster into a dot row representing the presence / absence of dot formation by determining the presence / absence of dot formation for each of the constituent pixels;
Second error diffusion means for diffusing the gradation error generated in each pixel constituting the head raster to undecided pixels around each pixel;
For the remaining rasters obtained by removing the head raster from the raster group, each of the remaining rasters is considered in consideration of the gradation error diffused from pixels that belong to the same raster group as the residual raster and for which dot formation has been determined. A residual raster converting means for converting the residual raster into a dot row in parallel with the process of converting the leading raster into the dot row by determining the presence or absence of dot formation of a pixel;
The first error diffusing unit and the second error diffusing unit store, in the first error storage unit, an error diffused to a pixel of a raster group different from the pixel for which the presence / absence of dot formation is determined, An image processing apparatus which is means for storing an error diffused to a pixel in the same raster group as the pixel for which the dot formation is determined or not, in a second error storage unit. The image processing apparatus according to claim 1 or 2 ,
The second data conversion means includes
Block forming means for dividing the raster group into a plurality of blocks by grouping a plurality of pixels constituting the raster group as a block composed of a plurality of pixels each having a predetermined number of pixels in the vertical and horizontal directions;
An image processing apparatus comprising: a block unit conversion unit that converts the rasters into N dot columns in parallel by determining whether or not the dots are formed in units of blocks. The image processing apparatus according to claim 4 ,
The second data conversion means performs the second image conversion process as a computer process for mounting a cache,
When there are a plurality of pixels for which the process of receiving the distribution of the error caused by the processing of other pixels in the block has been completed in the block-by-block conversion unit, An image processing apparatus that prioritizes selection of pixels that can use error distribution data stored in the cache. An image processing apparatus according to any one of claims 1 to 5 ,
The color image data is data in which gradation values of each color including at least cyan, magenta, and yellow are associated with each of the pixels,
The image processing apparatus, wherein the first color component includes at least a yellow color. An image processing apparatus according to any one of claims 1 to 5 ,
The color image data is data in which gradation values of each color including at least cyan, magenta, and yellow are associated with each of the pixels,
The image processing apparatus, wherein the second color component includes at least a cyan color or a magenta color. An image processing apparatus according to any one of claims 1 to 5 ,
The color image data is data in which gradation values of each color including at least cyan, magenta, yellow, and black are associated with each of the pixels,
The image processing apparatus, wherein the second color component includes at least a black color. An image processing apparatus according to any one of claims 1 to 5 ,
The color image data is data in which gradation values of the predetermined plurality of colors including at least a set of shades having substantially the same hue and different densities are associated with each of the pixels,
An image processing apparatus in which one color in the set of shades is selected as the first color component and the other color is selected as the second color component. By determining the presence or absence of dot formation for each of the pixels constituting the color image data for each color component, the color image data is converted into print data in an expression format based on the presence or absence of dot formation, and each color based on the print data A printing apparatus for expressing a color image on a print medium by forming dots of
Raster group generation means for generating a raster group by combining predetermined N (N is an integer of 2 or more) rasters that are adjacent to each other, which are rasters of pixel rows;
For the image data of the first color component selected in advance from the color components, the raster group is configured by comparing the gradation value of each pixel with a predetermined threshold value associated with each pixel. First data conversion means for performing first image conversion processing for converting the N rasters to be converted into dot rows in parallel, thereby converting the N rasters into first print data in an expression format depending on the presence or absence of dot formation; ,
For the image data of the second color component obtained by removing the first color component from each color component, the gradation error caused by the pixel determined by determining the presence / absence of dot formation is diffused to surrounding undetermined pixels. At the same time, the N rasters constituting the raster group are converted into dot rows in parallel while judging the presence / absence of dot formation so as to eliminate the gradation error diffused from the surrounding judged pixels. By doing so, based on the second data conversion means for performing the second image conversion processing for converting into the first print data of the expression format depending on the presence or absence of dot formation, the first print data and the second print data And dot forming means for forming dots of each color on the print medium,
The first and second data conversion means perform the first and second image conversion processing as processing of a computer that implements at least one of cache and pipeline processing ,
The first data conversion means includes
Threshold matrix storage means for storing a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged;
Target pixel selection means for selecting a target pixel for determining whether or not to form a dot;
A dot formation determination unit that reads out a threshold value corresponding to the target pixel from the threshold value matrix and compares the threshold value with the gradation value corresponding to the target pixel, thereby determining whether or not the target pixel is formed. When
With
The target pixel selecting means is means for selecting the target pixel in parallel from each raster constituting the raster group by sequentially selecting each pixel included in the raster group in a predetermined order.
The order of pixel selection in the target pixel selection means is up to two pixels that are continuously selected on the same raster.
Printing device. An image processing method for converting the color image data into data of each color in an expression format according to the presence or absence of dot formation by determining the presence or absence of dot formation for each pixel constituting each color image data,
Separating the received color image data into image data of a first color component preselected from the color components and image data of a second color component excluding the first color component; (A) and
For the image data of the first color component, an adjacent raster among rasters that are columns of pixels is determined by comparing the gradation value of each pixel with a predetermined threshold value associated with each pixel. A step (B) of performing a first image conversion process for converting the data into an expression format data according to the presence or absence of dot formation by converting N dots (N is an integer of 2 or more) in parallel to each other;
As for the image data of the second color component, the gradation error generated in the pixel determined by determining the presence / absence of dot formation is diffused to the surrounding undetermined pixels and diffused from the surrounding determined pixels. By converting the adjacent rasters into dot rows in parallel while determining the presence / absence of dot formation so as to eliminate the tone error, the data is converted into data in an expression format according to the presence / absence of dot formation. And (C) performing the second image conversion process,
In the steps (B) and (C), the first and second image conversion processes are performed as a process of a computer that implements at least one of a cache process and a pipeline process ,
The step (B)
Storing a threshold matrix in which a plurality of the thresholds are two-dimensionally arranged (b1);
A step (b2) of selecting a target pixel for determining whether or not to form a dot;
A step (b3) of determining whether or not a dot is formed for the target pixel by reading out a threshold corresponding to the target pixel from the threshold matrix and comparing it with the gradation value corresponding to the target pixel. When
With
The step (b2) is a step of selecting the target pixel in parallel from each raster constituting the raster group by sequentially selecting each pixel included in the raster group in a predetermined order.
The order of pixel selection in the step (b2) is up to two pixels that are continuously selected on the same raster.
Image processing method.

Images