JP2008199472A - Image processing apparatus and image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform dithering capable of achieving high gradation at low cost. <P>SOLUTION: A dithering part 14 applies matrix set with a plurality of thresholds in a gradation conversion processing part 15 to an image, and performs gradation conversion of representative gradation value of an image region corresponding to the matrix based on thresholds of the matrix. An error calculating part 16 calculates gradation error caused in the image region corresponding to the matrix before and after the gradation conversion. An error diffusing part 18 diffuses the gradation error calculated to the threshold of the peripheral matrix applied in the periphery of the image region corresponding to the matrix. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image processing apparatus and an image processing method for performing dither processing for reproducing halftones.

  2. Description of the Related Art Conventionally, in a printer or the like, as a method of reproducing a halftone, a dither processing technique is used in which a pixel value of an image is binarized or multi-valued to express a gradation using a matrix in which a plurality of threshold values are set. Yes. In dither processing, the gradations that can be expressed are limited by the size of the matrix, that is, the number of threshold values set in the matrix. On the other hand, there is a method called a super cell as one of the methods for realizing higher gradation. In the supercell, a plurality of small unit matrices are combined to form one large matrix. Since the number of thresholds increases, the number of gradations that can be expressed increases and an image with high gradation can be obtained.

  However, since the supercell method has a large number of thresholds, a large-capacity memory is required and the cost is high. Also, since supercells are periodically applied to an image as in the case of a normal matrix, moire may occur due to repeated patterns having the same threshold.

As another method, a representative gradation value is obtained for each pixel group, the number of dots corresponding to the representative gradation value is determined based on a predetermined correspondence relationship, and the gradation value indicated by the number of dots is There is a method of diffusing an error from a representative gradation value to pixel values of surrounding pixels (for example, see Patent Document 1). According to this method, it is possible to eliminate a gradation error from the original image caused by binarization or multi-value conversion, and the image reproducibility is improved.
JP 2005-303666 A

However, according to the method described in Patent Document 1, since the number of dots output by the representative gradation value is determined in advance, there is a limit to the number of gradations that can be expressed.
In the method described in Patent Document 1, the order of pixels in which dots are output in the pixel group is specified, and the position of the dot to be output is determined based on this order. In this case, the dots are output in the same order and at the same position, and it is highly possible that moire occurs due to the repetition. Further, when the determination is made by a plurality of ranks, a large-capacity memory is required to store the ranks, resulting in high costs.

  An object of the present invention is to perform dither processing that can realize high gradation at low cost.

The invention according to claim 1 is an image processing apparatus,
A gradation conversion processing unit that applies a matrix in which a plurality of threshold values are set to an image, and performs gradation conversion of a representative gradation value of an image area corresponding to the matrix based on the threshold value of the matrix;
Before and after the gradation conversion, an error calculation unit that calculates a gradation error generated in an image area corresponding to the matrix;
An error diffusion unit for diffusing the calculated gradation error to a threshold value of a peripheral matrix applied around an image area corresponding to the matrix;
It is characterized by providing.

The invention according to claim 2 is the image processing apparatus according to claim 1,
The plurality of threshold values set in the matrix are different from each other.

The invention according to claim 3 is the image processing apparatus according to claim 1 or 2,
The gradation conversion processing unit performs gradation conversion of the representative gradation value into a gradation value expressed in binary or multivalue based on a threshold value of the matrix.

The invention according to claim 4 is the image processing apparatus according to any one of claims 1 to 3,
The gradation conversion processing unit determines the representative gradation value based on a pixel value of each pixel constituting the image region to which the matrix is applied.

The invention according to claim 5 is the image processing apparatus according to any one of claims 1 to 4,
The gradation error is a difference between a representative gradation value before the gradation conversion is performed and a gradation value after the gradation conversion is performed.

The invention according to claim 6 is the image processing apparatus according to any one of claims 1 to 4,
The gradation error includes a representative gradation value before the gradation conversion is performed, and a gradation value obtained by a maximum threshold value in which dots are not output in each pixel of the image area by the gradation conversion. It is the difference of these.

The invention according to claim 7 is the image processing apparatus according to any one of claims 1 to 6,
The error diffusion unit weights a gradation error and adds the weighted error to a threshold value of the peripheral matrix.

The invention according to claim 8 is the image processing apparatus according to claim 7,
The error diffusion unit performs weighting of the gradation error and addition to the threshold value of the peripheral matrix using a diffusion coefficient pattern in which a threshold value position and a weighting coefficient for adding the gradation error in the peripheral matrix are predetermined. It is characterized by that.

The invention according to claim 9 is the image processing apparatus according to claim 8,
The diffusion coefficient pattern includes a plurality of diffusion coefficient patterns having different threshold positions and weighting coefficients for adding gradation errors in the peripheral matrix,
An operation unit for selecting a diffusion coefficient pattern to be used in the error diffusion unit from among the plurality of diffusion coefficient patterns is provided.

The invention according to claim 10 is an image processing method,
A gradation conversion processing step of applying a matrix in which a plurality of threshold values are set to an image, and performing gradation conversion of a representative gradation value of an image region corresponding to the matrix based on the threshold values of the matrix;
An error calculating step for calculating a tone error generated in an image region corresponding to the matrix before and after the tone conversion;
An error diffusion step of diffusing the calculated gradation error to a threshold value of a peripheral matrix applied around an image region corresponding to the matrix;
It is characterized by including.

The invention according to claim 11 is the image processing method according to claim 10,
The plurality of threshold values set in the matrix are different from each other.

The invention according to claim 12 is the image processing method according to claim 10 or 11,
In the gradation conversion processing step, the representative gradation value is converted into a gradation value expressed in binary or multi-value based on a threshold value of the matrix.

The invention according to claim 13 is the image processing method according to any one of claims 10 to 12,
In the gradation conversion processing step, the representative gradation value is determined based on a pixel value of each pixel constituting the image area to which the matrix is applied.

The invention according to claim 14 is the image processing method according to any one of claims 10 to 13,
The gradation error is a difference between a representative gradation value before the gradation conversion is performed and a gradation value after the gradation conversion is performed.

The invention according to claim 15 is the image processing method according to any one of claims 10 to 14,
The gradation error is a representative gradation value before the gradation conversion is performed and a gradation value obtained by a maximum threshold value at which dots are not output in each pixel of the image area by the gradation conversion. It is a difference.

The invention according to claim 16 is the image processing method according to any one of claims 10 to 15, wherein
In the error diffusion step, a gradation error is weighted and added to a threshold value of the peripheral matrix.

The invention according to claim 17 is the image processing method according to claim 16,
In the error diffusion step, the gradation error is weighted and added to the threshold value of the peripheral matrix using a diffusion coefficient pattern in which the threshold value position and weighting coefficient for adding the gradation error in the peripheral matrix are determined in advance. It is characterized by that.

The invention according to claim 18 is the image processing method according to claim 17,
The diffusion coefficient pattern includes a plurality of diffusion coefficient patterns having different threshold positions and weighting coefficients for adding gradation errors in the peripheral matrix,
In the error diffusion step, the gradation error is weighted and added to a threshold value of a peripheral matrix using a diffusion coefficient pattern selected via an operation unit among the plurality of diffusion coefficient patterns. .

  According to the invention described in claims 1, 5, 6, 10, 14, and 15, only the gradation error can be canceled by the entire image by diffusing the gradation error to the threshold value of the peripheral matrix. Instead, it is possible to reproduce an image with high gradation as in the supercell in which different threshold values are set in the same matrix unit. Further, since it is only necessary to sequentially change the threshold value of a certain basic matrix by error diffusion, it is only necessary to hold the threshold value of the basic matrix, and the same effect as a super cell can be obtained with a small memory capacity. it can. Therefore, it is possible to perform dither processing that can realize high gradation at low cost. Further, since the diffused gradation error varies depending on the image region to which the matrix is applied, the threshold value of the matrix can be changed at random, and image quality degradation such as moire caused by the repeated pattern of the threshold value can be avoided.

  According to the second and eleventh aspects of the present invention, it is possible to realize higher gradation by setting all the threshold values of the matrix to different values.

  According to the third and twelfth aspects of the present invention, binary or multi-level gradation expression is possible depending on the purpose.

  According to the inventions of claims 4 and 13, the average value of the pixel values of each pixel, the pixel value of a specific pixel among the pixels, and the like can be used as the representative gradation value, and the reproducibility of the original image good.

  According to the invention described in claims 7 to 9 and 16 to 18, a diffusion coefficient pattern according to the purpose can be selected from a plurality of diffusion coefficient patterns, and the degree of gradation error diffusion is adjusted. Can do. Thereby, the image quality can be adjusted.

First, the configuration will be described.
FIG. 1 shows a configuration of an MFP (Multi Function Peripheral) 100 according to the present embodiment. The MFP 100 is a multifunction machine having multiple functions such as a printer function, a copying function, and a scanner function. As shown in FIG. 1, the MFP 100, an image reading unit 20, an operation unit 30, a touch panel 40, a display unit 50, a printer. The unit 60 is provided.

Hereinafter, each part will be described.
The image reading unit 20 includes a light source, a CCD (Charge Coupled Device), an A / D converter, and the like, and reads an image of a document. At the time of image reading, the reflected light of the light scanned from the light source to the original is imaged and photoelectrically converted by the CCD to read the original image and generate an image signal (analog signal). Thereafter, the image signal is converted into digital image data by an A / D converter and output to the main body unit 10.

  The operation unit 30 includes a touch panel 40 and the like in addition to various function keys. When an operation is performed on each key or the touch panel 40, the operation unit 30 generates an operation signal corresponding to the operation and outputs the operation signal to the main body unit 10.

  The display unit 50 includes a display display that is integrated with the touch panel 40. The display unit 50 displays various operation screens on the display display according to the control of the control unit 2 of the main body unit 10.

  The printer unit 60 prints out an image to be printed on a print sheet. For example, when the electrophotographic method is adopted, the printer unit 60 includes an exposure unit having a photosensitive drum, a developing unit for attaching toner, a fixing unit for performing toner fixing processing, and the like. At the time of printing, an electrostatic latent image based on image data is written on the photosensitive drum in the exposure unit, and toner is attached to the photosensitive drum in the developing unit to form a toner image. This toner image is transferred onto the print paper conveyed from the paper feed tray. Thereafter, the print sheet is conveyed to a fixing unit, subjected to a fixing process, and output to a designated discharge destination tray. The printing method may be a method other than the electrophotographic method.

Next, the main body 10 will be described.
As shown in FIG. 1, the main body unit 10 includes an image processing apparatus 1, a control unit 2, a storage unit 3, a DRAM (Dynamic Random Access Memory) control unit 4, and a DRAM 5.

  The control unit 2 includes a CPU (Central Processing Unit), a RAM (Random Access Memory), and the like, and performs overall control of operations of each unit of the MFP 100 according to various control programs stored in the storage unit 3. Various calculations are also performed. For example, in accordance with an operation signal input from the operation unit 30, the copy, printer, and scanner modes are switched, a document is read by the image reading unit 20, stored in the DRAM 5, and image data output to the printer unit 60. I do.

  The storage unit 3 stores various programs and files and data necessary for executing the programs. For example, basic matrix data used in a dither processing unit of the image processing apparatus 1 described later, diffusion coefficient pattern data, and the like are stored.

The DRAM control unit 4 controls input / output of image data to / from the DRAM 5 according to the control of the control unit 2.
The DRAM 5 includes a storage memory area and a page memory area. The storage memory area is a memory for storing image data, and the page memory area is a memory for temporarily storing image data to be printed before printing.

Next, the image processing apparatus 1 according to the present invention will be described.
The image processing apparatus 1 performs various types of image processing on image data to be printed.
As shown in FIG. 2, the image processing apparatus 1 includes a color conversion unit 11, a γ correction unit 13, and a dither processing unit 14.

  The color conversion unit 11 converts the image data input from the image reading unit 20 into image data corresponding to a color material that can be output by the MFP 100. For example, when R, G, and B color image data is input from the image reading unit 20, these color image data are converted into color material Y, M, C, and K color image data that can be output by the MFP 100. Convert and output.

  The γ correction unit 13 performs γ correction processing by converting input image data using an LUT (lookup table) prepared in advance for γ correction. The image data that has been subjected to the γ correction processing is output to the dither processing unit 14.

The dither processing unit 14 performs dither processing on the input image data to reproduce halftones. At the time of dither processing, a matrix in which a plurality of threshold values are set in advance in a matrix is read from the storage unit 3 and used for processing.
As shown in FIG. 2, the dither processing unit 14 includes a gradation conversion processing unit 15, an error calculation unit 16, an error memory 17, and an error diffusion unit 18. Hereinafter, the dither processing performed by these units will be described with reference to the flowchart shown in FIG.

  The gradation conversion processing unit 15 applies a matrix to the input image, binarizes or multi-values each pixel value of the image area corresponding to the matrix, and converts the gradation value of the image area to binary or multi-value. Is converted into a gradation value expressed by the pixel value. Hereinafter, a binary gradation expression may be referred to as a binary expression, and a multi-value gradation expression that is greater than two values may be referred to as a multi-value expression.

FIG. 4 shows an example of a basic matrix used for binarization.
FIG. 4A shows an example of the matrix m1 used in the AM system, and FIG. 4B shows an example of the matrix m2 used in the FM system. The AM method reproduces halftones according to the size of halftone dots formed by a matrix, whereas the FM method reproduces halftones according to the density of adjacent halftone dots with the size of the halftone dots themselves being constant. It is.

Each of the matrices m1 and m2 includes a plurality of cells (one cell corresponds to one pixel), and a threshold value (a number displayed in the cell in FIG. 4) is set for each cell. The threshold is set on the assumption that the pixel value 255 is assigned to white and the pixel value 0 is assigned to black.
Based on the result of comparing the threshold value and each pixel value of the image area corresponding to the matrix, each pixel value of the image area is binarized. That is, the presence / absence of dot output is determined for each pixel. With this set of dots, a halftone dot is formed for each image area to which the matrix is applied. That is, the gradation of the image area is expressed by the output of the dot, and the gradation value of the image area determines the density of the halftone dots. Therefore, the higher the number of gradations that can be taken as the gradation value of the image area, the higher the gradation can be realized. Note that the threshold values set in the matrices m1 and m2 are preferably different values. This is because, by setting all values to be different, it is possible to increase the number of gradations that can be expressed by the matrices m1 and m2, and to realize a high gradation.

  When the image data to be processed is input (step S1), the gradation conversion processing unit 15 applies a matrix to the input image (step S2). Since the matrix is configured to be complementary between adjacent matrices, the entire image can be scanned without omission by sequentially shifting the application position of the matrix. Hereinafter, gradation conversion processing is performed in the applied matrix. The matrix being processed is referred to as a target matrix, the matrix around the target matrix, and the matrix to be processed thereafter is referred to as a peripheral matrix.

  The gradation conversion processing unit 15 determines a representative gradation value of the image area corresponding to the attention matrix (Step S3). The representative gradation value is an average value obtained by averaging the pixel values of the pixels in the image area corresponding to the matrix. Alternatively, the position of a specific pixel to be used as the representative gradation value in the matrix is set in advance, and the pixel corresponding to the setting position is determined as the specific pixel in the image area corresponding to the matrix, and the pixel value is set as the representative gradation. It may be a value. After the determination, the gradation conversion processing unit 15 replaces each pixel value in the image area with a representative gradation value.

This will be described with reference to FIG.
For example, when the matrix m1 shown in FIG. 4A is applied to the input image, and the image area corresponding to the matrix m1 is the image area g11 shown in FIG. An average value of pixel values of each pixel constituting g1 is obtained, and this is set as a representative gradation value of the image region g1. Since the average value of each pixel value in the image region g1 is 100, the gradation conversion processing unit 15 replaces the representative gradation value of 100 with the pixel value of each pixel. The result of the replacement is an image region g12 shown in FIG.

Next, the gradation conversion processing unit 15 compares each pixel value in the image area, that is, the replaced representative gradation value, with each threshold value set in the matrix, and performs gradation conversion based on the comparison result. (Step S4). Referring to the example of FIG. 5, in the case of binarization, if the representative gradation value is equal to or higher than the threshold value of the matrix m1, the pixel value is converted to a pixel value 255 (white pixel), and if it is less than the threshold value, the pixel value is 0 (black Pixel). Since the pixel value 255 is a dot non-output and the pixel value 0 is a dot output, the output state of the dot is as shown in an image region g13 shown in FIG.
Note that the threshold value of the matrix may be changed by an error diffusion unit 18 described later. In this case, gradation conversion is performed as described above using a matrix whose threshold value has been changed.

  When multi-value processing is performed, a method different from binarization is used. In the case of binarization, one threshold value is set for one cell in the matrices m1 and m2, whereas in the case of multilevel conversion, two independent threshold values are set for one cell in the matrix m1 and m2. Set. Then, multi-leveling is performed using these two threshold values. Specifically, a conversion function defined by two threshold values TH1 [n] and TH2 [n] (n is a number indicating a cell) set in each cell is created, and a pixel value to be compared with this conversion function Is determined as a pixel value after multi-value conversion. For example, in the case of a cell of n = 2, the conversion function shown in FIG. 6 is created with the thresholds TH1 [2] and TH2 [2]. Therefore, when the representative gradation value of the pixel corresponding to the cell of n = 2 is 100, the pixel value 125 is output by the conversion function as shown in FIG. In actual processing, an LUT is created in advance based on a conversion function, and a pixel value after multi-value conversion is obtained using this LUT.

  For example, in the case of threshold values TH1 [n] and TH2 [n] determined when the slope of the threshold function is 255/4, the output is expressed in five values. In the case of quinarization, when TH1 [n] = 127 and TH2 [n] = 131 are set in a certain cell in the matrices m1 and m2, the pixel value of the image corresponding to the cell is set to 127 by gradation conversion. If it is smaller, 0 value is output from the LUT. If the value is 127 or more and less than 128, 64 value is output. If it is 128 or more and less than 129, 128 value is output. .

FIG. 7 is a diagram illustrating an example of performing gradation conversion by multi-value expression using the matrix m2.
An image area g21 shown in FIG. 7 is obtained by replacing the gradation value of each pixel with the representative gradation value 100 in the image area corresponding to the matrix m2. The gradation conversion processing unit 15 multi-values the pixel value (representative gradation value) of the image region g21 using two threshold values set in each cell of the matrix m2. The pixel value 0 outputs one dot, the pixel value 255 does not output a dot, and when the pixel value is 0 or more and less than 255, a dot corresponding to the pixel value is output. The output state of the dot is shown in FIG. It looks like the image area g22.

  The error calculation unit 16 calculates a tone error that occurs before and after tone conversion in the image region corresponding to the attention matrix (step S5). The gradation error is a difference between the representative gradation value before gradation conversion and the gradation value of the image area after gradation conversion. For example, as shown in FIG. 8, a case will be described in which an image region g11 having a representative gradation value of 100 is changed to a binary representation image region g13 by gradation conversion. The maximum gradation of an image (8-bit data) is 255, and this does not change before and after gradation conversion. Since all 15 pixels of the matrix m1 are white (255 value), 10 pixels out of 15 pixels are black (0 value), so the gradation expressed by the image region g13 is (15-10). ) / 15 × 255 = 5/15 × 255 = 85 gradations. Therefore, the gradation error is a difference of -15 between the gradation value 85 after gradation conversion and the representative gradation value 100 before gradation conversion.

That is, (number of white (255 values) pixels in matrix m1 after gradation conversion) / (number of white pixels when all pixels in matrix m1 are white pixels) × (maximum gradation of image) ) To calculate the gradation after gradation conversion.
In the case of multi-value expression, for example, when a 15-pixel matrix that performs quinarization converts 4 pixels into black (0 value) and 1 pixel into an intermediate value (128 values), an intermediate value These pixels can be calculated as 10.5 / 15 × 255 = 179 (rounded off after the decimal point) gradation from the above equation, with black pixels converted to 0.5 pixels.

  Note that the tone error may be calculated based on the maximum threshold value among the threshold values for which dots are not output for each pixel by tone conversion. In this case, the difference between the maximum threshold value and the representative gradation value among the threshold values in which dots are not output, that is, the pixel value is set to 1 by gradation conversion, is the gradation error. Here, the maximum threshold value refers to a threshold value having the smallest difference from the pixel value compared with the threshold value among the threshold values having a pixel value of 1, and does not refer to the size of the threshold value itself. In the example of binary expression shown in FIG. 5, among the 15 pixels corresponding to the matrix m1, 5 pixels are not output as dots (the pixel value is 1), so this dot is not output. A difference of −15 between the maximum threshold value 85 of the threshold values 17, 51, 34, 68, and 85 and the representative gradation value 100 is a gradation error. The same applies to the case of multi-value expression.

The error memory 17 stores the gradation error value calculated by the error calculation unit 16.
The error diffusion unit 18 uses the diffusion coefficient pattern to diffuse the gradation error stored in the error memory 17 to the threshold value of the peripheral matrix applied in the peripheral image region of the matrix of interest (step S6).
The diffusion coefficient pattern designates the position of the peripheral matrix for diffusing the gradation error and the weighting coefficient for weighting performed when diffusing the gradation error. There are a plurality of different diffusion coefficient patterns.

FIG. 9 shows an example of the diffusion coefficient pattern.
The diffusion coefficient patterns P1 to P4 shown in FIG. 9 are represented by a plurality of blocks, and these blocks represent a matrix. The block indicated by “*” indicates the position of the attention matrix, and the other blocks indicate peripheral matrices that diffuse gradation errors. The number displayed in the block of the peripheral matrix is a weighting coefficient for diffusing gradation errors in the peripheral matrix. The peripheral matrix designated in the diffusion coefficient patterns P1 to P4 is designed to designate a matrix that is located around the matrix of interest and that has not yet been subjected to gradation conversion.
The data of these diffusion coefficient patterns P1 to P4 is stored in the storage unit 3, and the error diffusion unit 18 reads it from the storage unit 3 and uses it for processing as necessary.

  Of the plurality of diffusion coefficient patterns P1 to P4, the user can select diffusion coefficient patterns P1 to P4 used by the error diffusion unit 18 for processing. That is, when the user selects and operates any one of the diffusion coefficient patterns P1 to P4 via the operation unit 30, the error diffusion unit 18 reads the selected diffusion coefficient patterns P1 to P4 and uses them for processing. The diffusion coefficient patterns P1 and P2 have a smaller number of peripheral matrices than the diffusion coefficient patterns P3 and P4, and their positions are also adjacent to the target matrix. That is, since the range in which the gradation error is diffused is narrow, the reproducibility is improved when the image to be processed includes characters, line drawings, and the like.

  On the other hand, since the diffusion coefficient patterns P3 and P4 have a wide range of diffusing gradation errors, the halftone reproducibility is improved. Further, depending on not only the position and number of the peripheral matrix but also the setting of the weighting coefficient, a large amount of gradation error is diffused to the peripheral matrix close to the target matrix, or evenly distributed to all the peripheral matrices. Therefore, the reproducibility of the image changes. As described above, since the effects obtained by the diffusion coefficient patterns P1 to P4 to be used are different, the user can adjust the image quality by selecting the diffusion coefficient patterns P1 to P4 according to the purpose.

  First, the error diffusion unit 18 reads gradation error data from the error memory 17 and reads the diffusion coefficient patterns P1 to P4 selected by the user. Then, the weighting coefficients designated by the diffusion coefficient patterns P1 to P4 are read out, and the gradation error is multiplied to perform weighting. Further, a peripheral matrix that diffuses gradation errors based on the diffusion coefficient patterns P1 to P4, that is, a matrix in a region where gradation conversion has not yet been performed, is determined, and each threshold value of the peripheral matrix is weighted by the weighting coefficient. Add the gray level error.

A specific example in which error diffusion is performed using the diffusion coefficient pattern P2 is shown in FIG. FIG. 10 shows an example in which the gradation error −15 is diffused.
In FIG. 10, reference numeral M1 indicates a matrix of interest. Reference numerals M2 to M5 indicate a peripheral matrix designated by the diffusion coefficient pattern P2. In the diffusion coefficient pattern P2, the weighting coefficients of the peripheral matrices M2, M3, M4, and M5 are specified as 2/8, 3/8, 2/8, and 1/8, respectively.

For example, when the gradation error to be diffused is −15, the weighted gradation errors diffused to the peripheral matrices M2, M3, M4, and M5 are −4, −6, −4, and −2, respectively (actual The calculated values are −3.75, −5.625, −3.75, and −1.875, but the numbers after the decimal point are rounded up). The result of adding the weighted gradation error to each threshold value of the basic matrix m1 and applying it to the image is the peripheral matrices M2 to M5 shown in FIG.
As a result of the gradation error being diffused in this way, when the gradation conversion processing unit 15 performs gradation conversion at the position of the peripheral matrix, gradation conversion is performed using the matrix m1 whose threshold has been changed. .

  As described above, according to the present embodiment, the matrix is sequentially applied to the image, and the representative gradation value of the image area corresponding to the attention matrix is calculated. The representative gradation value is subjected to gradation conversion based on the threshold value of the matrix of interest, and the gradation error generated before and after the gradation conversion is diffused to the peripheral matrix based on the diffusion coefficient pattern. By diffusing the gradation error, it is possible to cancel the gradation error caused by the gradation conversion in the entire image. Therefore, the reproducibility of the image is improved.

  Further, the threshold value of the basic matrix is sequentially changed by diffusing the gradation error, and gradation conversion is performed using the changed matrix. As a result, it is possible to perform the same processing as when applying a supercell in which different threshold values are set in the same matrix unit, and it is possible to realize de-processing with high gradation. In the case of a supercell, a large number of threshold values are set and a large-capacity memory is required to store these threshold values. However, according to this embodiment, threshold values can be stored only for the basic matrix. Therefore, the same effect as that of the supercell can be obtained with a small memory capacity.

  In the case of a supercell, moire may occur due to the repeated pattern of the supercell threshold as a result of repeatedly applying the supercell to the image. However, according to the present embodiment, since the diffused gradation error varies depending on the image area to which the matrix is applied, the matrix threshold value is randomly changed. Thereby, it can avoid that the pattern of the same threshold value is applied repeatedly, and generation | occurrence | production of a moire can be suppressed.

  In the matrix, different threshold values are set for all the cells constituting the matrix. As a result, the number of gradations that can be expressed by the matrix can be increased, and the gradation is further improved.

  In the gradation conversion process, it is possible to convert to a gradation value expressed by binary or multivalue. As a result, binary or multi-level gradation expression is possible depending on the purpose.

  Further, in the image area corresponding to the matrix, instead of performing gradation conversion using the pixel value of each pixel in the image area, the representative gradation value in the image area is calculated and this representative gradation value is used. Gradation conversion and gradation error calculation. Therefore, it is not necessary to perform calculation using different pixel values for each pixel, and gradation conversion processing becomes efficient and gradation error calculation is facilitated.

  Further, the position of the threshold for diffusing the gradation error and the degree of weighting are determined based on the diffusion coefficient pattern. There are a plurality of types of diffusion coefficient patterns, and it is possible to select which one to apply via the operation unit. Therefore, a diffusion coefficient pattern according to the purpose can be selected, and image quality can be adjusted by adjusting the degree of gradation error diffusion.

2 is a diagram illustrating a configuration of an MFP. FIG. It is a figure which shows the structure of the image processing apparatus which concerns on this embodiment. It is a flowchart explaining the flow of a process in an image processing apparatus. (A) It is a figure which shows the matrix example of AM system. (B) It is a figure which shows the matrix example of FM system. It is a figure explaining the gradation conversion process in the case of carrying out binary expression using the threshold value of a matrix. It is a figure which shows the conversion type | formula in the case of gradation conversion by two threshold values. It is a figure explaining the gradation conversion process in the case of multi-value expression using two threshold values. It is a figure explaining the calculation method of a gradation error. It is a figure which shows the example of a spreading | diffusion coefficient pattern. It is a figure which shows the result of having diffused the gradation error to the threshold value of the surrounding matrix.

Explanation of symbols

100 MFP
DESCRIPTION OF SYMBOLS 1 Image processing apparatus 14 Dither processing part 15 Gradation conversion processing part 16 Error calculation part 17 Error memory 18 Error diffusion part

Claims (18)

A gradation conversion processing unit that applies a matrix in which a plurality of threshold values are set to an image, and performs gradation conversion of a representative gradation value of an image area corresponding to the matrix based on the threshold value of the matrix;
Before and after the gradation conversion, an error calculation unit that calculates a gradation error generated in an image area corresponding to the matrix;
An error diffusion unit for diffusing the calculated gradation error to a threshold value of a peripheral matrix applied around an image area corresponding to the matrix;
An image processing apparatus comprising:   The image processing apparatus according to claim 1, wherein the plurality of threshold values set in the matrix are different from each other.   The gradation conversion processing unit performs gradation conversion of the representative gradation value into a gradation value expressed in binary or multivalue based on a threshold value of the matrix. Image processing apparatus.   The said gradation conversion process part determines the said representative gradation value based on the pixel value of each pixel which comprises the image area to which the said matrix was applied. An image processing apparatus according to 1.   5. The gradation error is a difference between a representative gradation value before the gradation conversion is performed and a gradation value after the gradation conversion is performed. The image processing apparatus according to any one of the above.   The gradation error is a difference between a representative gradation value before the gradation conversion is performed and a maximum threshold value among the threshold values in which dots are not output in each pixel of the image area by the gradation conversion. The image processing apparatus according to claim 1, wherein the image processing apparatus is provided.   The image processing apparatus according to claim 1, wherein the error diffusion unit weights a gradation error and adds the weighted error to a threshold value of the peripheral matrix.   The error diffusion unit performs weighting of the gradation error and addition to the threshold value of the peripheral matrix using a diffusion coefficient pattern in which a threshold value position and a weighting coefficient for adding the gradation error in the peripheral matrix are predetermined. The image processing apparatus according to claim 7. The diffusion coefficient pattern includes a plurality of diffusion coefficient patterns having different threshold positions and weighting coefficients for adding gradation errors in the peripheral matrix,
The image processing apparatus according to claim 8, further comprising an operation unit for selecting a diffusion coefficient pattern to be used in the error diffusion unit from among the plurality of diffusion coefficient patterns. A gradation conversion processing step of applying a matrix in which a plurality of threshold values are set to an image, and performing gradation conversion of a representative gradation value of an image region corresponding to the matrix based on the threshold values of the matrix;
An error calculating step for calculating a tone error generated in an image region corresponding to the matrix before and after the tone conversion;
An error diffusion step of diffusing the calculated gradation error to a threshold value of a peripheral matrix applied around an image region corresponding to the matrix;
An image processing method comprising:   The image processing method according to claim 10, wherein the plurality of threshold values set in the matrix are different from each other.   12. The gradation conversion process step, wherein the representative gradation value is converted into a gradation value expressed in binary or multi-value based on a threshold value of the matrix. Image processing method.   13. The representative gradation value is determined in the gradation conversion processing step based on a pixel value of each pixel constituting an image area to which the matrix is applied. An image processing method described in 1.   14. The gradation error is a difference between a representative gradation value before the gradation conversion is performed and a gradation value after the gradation conversion is performed. The image processing method according to any one of the above.   The gradation error is a difference between a representative gradation value before the gradation conversion is performed and a maximum threshold value among the threshold values in which dots are not output in each pixel of the image area by the gradation conversion. The image processing method according to claim 10, wherein the image processing method is provided.   The image processing method according to claim 10, wherein in the error diffusion step, a gradation error is weighted and added to a threshold value of the peripheral matrix.   In the error diffusion step, the gradation error is weighted and added to the threshold value of the peripheral matrix using a diffusion coefficient pattern in which the threshold value position and weighting coefficient for adding the gradation error in the peripheral matrix are determined in advance. The image processing method according to claim 16. The diffusion coefficient pattern includes a plurality of diffusion coefficient patterns having different threshold positions and weighting coefficients for adding gradation errors in the peripheral matrix,
In the error diffusion step, the gradation error is weighted and added to a threshold value of a peripheral matrix using a diffusion coefficient pattern selected via an operation unit among the plurality of diffusion coefficient patterns. The image processing method according to claim 17.

Images