JP7123752B2 - Image processing apparatus, image forming apparatus, image processing method, image processing program, and recording medium - Google Patents

Description

The present invention relates to an image processing apparatus, an image processing method, and an image processing program, and more particularly to a method of determining halftone frequency based on image data in order to perform appropriate image processing on the image data.

In addition to conventional analog image forming devices such as electronic copiers, digital types are becoming more popular, and with the advancement of digital image processing technology, full-color digital copiers that reproduce color images with high image quality are becoming popular. It has been commercialized.

2. Description of the Related Art In a digital copier, the number of halftone dots of a document (halftone dot ruling) is determined from the image data of the document read by the scanner of the image reading unit, and image processing is performed according to the determination result.

In Patent Document 1, G image data among RGB data is input, and whether or not the central pixel is a peak pixel in a matrix of a predetermined size (eg, 3×3 pixels) is detected, and a predetermined area (eg, 20×9 pixels) is detected. If the number of peak pixels is large in the range, it is judged as a high halftone frequency, and if it is small, it is judged as a low halftone frequency.

Japanese Patent Application Laid-Open No. 2008-078830 (published on April 3, 2008) Japanese Patent Application Laid-Open No. 2002-232708 (published on August 16, 2002) Japanese Patent Application Laid-Open No. 2000-354167 (published on December 19, 2000)

However, if the number of lines is determined only by the feature value that indicates the frequency of density change between adjacent pixels, such as the peak pixel number, paper quality, paper density, print quality, presence or absence of show-through, etc. will affect the number of inversions. do. For example, even if the number of lines is the same, if the paper quality is poor, the print quality is poor due to the paper quality, the paper is not pure white but a little dark, or there is show-through, etc., each halftone dot of the document will be different. It tends to be read blurry and the number of inversions is small. In this case, the document is likely to be determined to be a document with a low number of lines.

Further, even if the number of screen lines is the same, the tendencies of the number of pixel value inversions differ between a single-color halftone dot and a mixed-color halftone dot. For example, when the image data of a document read by a reading device is RGB data, the G image data is theoretically composed only of magenta, which is complementary to magenta. Cyan, yellow, and black halftone dot components are also mixed. Therefore, the determination result changes depending on the configuration of mixed colors of the halftone dots of the document.

The present invention has been made in view of such a situation, and it is possible to add not only a feature quantity indicating the frequency of grayscale change between adjacent pixels but also a feature quantity indicating the degree of grayscale change and a density distribution to the feature quantity. It is an object of the present invention to provide an image processing apparatus or the like capable of increasing the accuracy of line number determination, performing image processing according to the determination result, and outputting an image in which moiré is difficult to perceive while maintaining sharpness. do.

In order to solve the above-described problems, the image processing device of the present invention includes:
an image data input unit for scanning an original and inputting it as image data;
The document based on the image data feature amount including a first feature amount indicating frequency of grayscale change between adjacent pixels of the image data and a second feature amount indicating the degree of grayscale change of the image data. a determination unit that determines the screen frequency of halftone dots of
an image processing unit that performs image processing on the image data according to the determination result of the determination unit;
characterized by comprising

The image processing method of the present invention is
an image data input step of scanning a document and inputting it as image data;
The document based on the image data feature amount including a first feature amount indicating frequency of grayscale change between adjacent pixels of the image data and a second feature amount indicating the degree of grayscale change of the image data. a judgment step for judging the number of halftone dots of
an image processing step of performing image processing on the image data according to the determination result of the determination step;
characterized by comprising

The image processing program of the present invention is
In the processor of the image processing device,
an image data input step of scanning a document and inputting it as image data;
The document based on the image data feature amount including a first feature amount indicating frequency of grayscale change between adjacent pixels of the image data and a second feature amount indicating the degree of grayscale change of the image data. a judgment step for judging the number of halftone dots of
an image processing step of performing image processing on the image data according to the determination result of the determination step;
is characterized by executing

The computer-readable recording medium of the present invention is
In the processor of the image processing device,
an image data input step of scanning a document and inputting it as image data;
The document based on the image data feature amount including a first feature amount indicating frequency of grayscale change between adjacent pixels of the image data and a second feature amount indicating the degree of grayscale change of the image data. a judgment step for judging the number of halftone dots of
an image processing step of performing image processing on the image data according to the determination result of the determination step;
An image processing program for executing is recorded.

According to the image processing apparatus and the like of the present invention, by adding a feature amount indicating the degree of shading change to the parameters for determining the number of lines, it is possible to determine whether the paper quality, paper density, print quality, presence or absence of show-through, etc. The result of number judgment is less likely to be influenced. Therefore, by determining the number of rulings based only on the feature value that indicates the frequency of grayscale changes between adjacent pixels, it is possible to determine the number of rulings based on factors such as paper quality, paper density, print quality, the presence or absence of show-through, and the composition of mixed colors in halftone dots. It is possible to prevent the result of

1 is a block diagram showing a schematic configuration of an image forming apparatus according to an embodiment; FIG. 3 is a block diagram showing the configuration of a document feature determination section in the embodiment; FIG. 4 is a flow chart showing the flow of processing in a document feature determination unit of the embodiment; FIG. 10 is an explanatory diagram showing a method of setting an exclusion area excluded from histogram generation targets in document image data; FIG. 10 is a diagram showing an operation example of determination processing for classifying halftone dot rulings in area pixel determination processing; It is an example of a filter applied to a halftone dot area. 2 is a block diagram showing a schematic configuration of a portion related to a scan output function of the image forming apparatus according to the embodiment; FIG. FIG. 10 is a diagram showing an example of an operation panel for setting thresholds for line number determination;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, the present invention will be specifically described based on the drawings showing the embodiments thereof.

FIG. 1 is a block diagram showing a schematic configuration of an image forming apparatus 1 according to this embodiment. In this embodiment, a case where the image forming apparatus 1 is an electrophotographic digital color copying machine will be described, but the configuration of the image forming apparatus 1 is not limited to this. Any device may be used as long as it has a function of performing processing.

The image forming apparatus 1 includes a color image input device 110, a color image processing device 120, a color image output device 130, an operation panel 140, and a control section 100, as shown in FIG. Further, the color image processing apparatus 120 includes an A/D (analog/digital) conversion unit 11, a shading correction unit 12, a document feature determination unit 13, an input tone correction unit 14, an area separation processing unit 15, a color correction unit 16, A black generation undercolor removal unit 17 , a spatial filter processing unit 18 , an output gradation correction unit 19 , and a gradation reproduction processing unit 20 are provided.

The color image input device 110 inputs image data of a document by reading the document and generating image data of the document. Then, the generated (input) image data is output to the color image processing device 120 . The configuration of the color image input device 110 is not particularly limited as long as it can generate image data of a document, and various conventionally known image input devices can be used. For example, it has a scanner unit (not shown) equipped with a CCD (Charge Coupled Device), reads the reflected light image from the original with the CCD, and converts it into RGB (R: red, G: green, B: blue) analog signals. You may use the thing of the structure which produces|generates the image data which consists of.

Color image processing device 120 converts image data input from color image input device 110 into A/D conversion unit 11, shading correction unit 12, document feature determination unit 13, input gradation correction unit 14, and area separation processing unit 15. , color correction unit 16, black generation undercolor removal unit 17, spatial filter processing unit 18, output gradation correction unit 19, and gradation reproduction processing unit 20, and finally CMYK (C: cyan, M: magenta , Y: yellow, and K: black) is output to the color image output device 130 .

The A/D converter 11 converts image data of analog signals corresponding to RGB to image data of digital signals.

The shading correction unit 12 corrects various distortions occurring in the illumination system, imaging system, and imaging system of the color image input device 110 with respect to the image data of digital signals corresponding to RGB sent from the A/D conversion unit 11. process to remove the Also, the shading correction unit 12 adjusts the color balance.

The document feature determination unit 13 converts the RGB signals (RGB reflectance signals) from which various distortions have been removed and the color balance has been adjusted by the shading correction unit 12 into density signals. Further, the document feature determination unit 13 performs (i) document type determination processing for determining whether the document is a text document, a printed photo document, or a text printed photo document in which text and printed photos are mixed, and ( ii) Background determination processing is performed to determine whether or not the document should be subjected to background removal processing. In the document type determination process, screen frequency determination processing is performed to determine the frequency of halftone dots printed on the document. Then, the document characteristic determination section 13 generates a document characteristic determination signal indicating the result of the document type determination processing, and the input gradation correction section 14, the color correction section 16, the black generation undercolor removal section 17, and the spatial filter processing section 18. , and to the tone reproduction processing unit 20 . The input gradation correction unit 14, the color correction unit 16, the black generation undercolor removal unit 17, the spatial filter processing unit 18, and the gradation reproduction processing unit 20 perform image processing on the image data according to the document feature determination signal. . The details of the document feature determination unit 13 will be described later.

The input gradation correction unit 14 removes background density and performs image quality adjustment processing such as contrast.

In other words, the background of the document (the paper on which characters are printed) has various colors. may be used excessively. For example, when forming an image in monochrome mode on a document with a yellowish background such as straw paper or a document with a colored background, if the background color is not treated as white, the white background (underground) will not be printed on the recording material when the image is output. is used. For this reason, in order to suppress the use of excess recording material, the input tone correction unit 14 performs background density removal processing (that is, processing to change the density of the background to white), and reduces the contrast. Perform adjustment processing.

Further, the input gradation correction unit 14 converts image data composed of RGB density signals (hereinafter, unless otherwise specified, RGB signals represent density signals (signals representing pixel values)) to a region separation processing unit. Output to 15.

The area separation processing unit 15 divides each pixel in the image data output from the input tone correction unit 14 into a background area, a photograph area (continuous tone area), a character area, and a halftone dot area based on the RGB density signals. Performs region separation processing to determine which of the regions belongs to. Based on the results of the segmentation processing, the segmentation processing unit 15 generates a segmentation identification signal indicating which region each pixel belongs to, and the color correction unit 16, the black generation undercolor removal unit 17, the space Along with outputting to the filter processing unit 18 and the gradation reproduction processing unit 20, the RGB signals input from the input gradation correction unit 14 are directly output to the subsequent color correction unit 16. FIG.

The color correction unit 16 converts the RGB signals input from the segmentation processing unit 15 into CMY signals according to the segmentation signal input from the segmentation processing unit 15 . In addition, the color correction unit 16 performs processing for removing color turbidity based on spectral characteristics of CMY color materials including unnecessary absorption components in order to achieve faithful color reproduction.

The black generation undercolor removal unit 17 performs black generation processing to generate a black (K) signal from the CMY signals after color correction by the color correction unit 16, and subtracts the K signal obtained by the black generation processing from the original CMY signals. and undercolor removal processing for generating new CMY signals. As a result, the CMY signals are converted into CMYK four-color signals (hereinafter referred to as CMYK signals).

For example, the black generation and undercolor removal unit 17 performs black generation using skeleton black as black generation processing. In the black generation process using this skeleton black, the input/output characteristic of the skeleton curve is y=f(x), and the densities corresponding to the C signal, M signal, and Y signal of the input CMY signals are C, M, and Y, respectively. Y, C', M', Y', and K' respectively correspond to the C signal, M signal, Y signal, and K signal of the output CMYK signals, and α is the UCR (Under Color Removal) rate. Assuming (0<α<1),
K′=f{min(C,M,Y)}
C'=C-αK'
M′=M−αK′
Y′=Y−αK′
converts CMY three-color density signals into CMYK four-color density signals.

The spatial filter processing unit 18 corrects the spatial frequency characteristics of the image data of the CMYK signals input from the black generation and undercolor removal unit 17 to prevent blurring and graininess deterioration of the image data. Spatial filtering is performed using a digital filter preset for each type of region according to the region identification signal input from the separation processing unit 15 . Details of the spatial filtering unit 18 will be described later.

The output gradation correction unit 19 performs output gradation correction processing based on the output characteristics of the color image output device 130 on the image data of the CMYK signals input from the spatial filter processing unit 18 .

The gradation reproduction processing unit 20 predetermines the image data of the CMYK signals input from the output gradation correction unit 19 for each region type in accordance with the region identification signal input from the region separation processing unit 15. the specified processing.

For example, regarding the regions separated into character regions by the region separation processing unit 15, in order to improve the reproducibility of black characters or color characters in particular, the spatial filter processing unit 18 increases the amount of high-frequency enhancement by sharpness enhancement processing. A tone reproduction processing unit 20 performs binarization or multi-value processing on a high-resolution screen suitable for reproduction of high frequencies.

In addition, for the regions separated into halftone dot regions by the region separation processor 15, the spatial filter processor 18 performs low-pass filter processing for removing input halftone dot components, and the gradation reproduction processor 20 performs the respective Gradation reproduction processing (halftone generation) is performed so that the gradation of pixels can be reproduced.

In addition, regarding the area separated into a photographic area (continuous tone area) by the area separation processing unit 15, the tone reproduction processing unit 20 performs binarization or multi-value processing on a screen that emphasizes tone reproduction. conduct.

The image data output from the gradation reproduction processing section 20, that is, the image data that has undergone each of the processes described above, is temporarily stored in a storage section (not shown), read out at a predetermined timing, and supplied to the color image output apparatus. 130.

The color image output device 130 outputs an image corresponding to image data to a recording medium (for example, a sheet such as paper). The color image output device 130 forms an image on a recording medium using recording materials of two or more colors (four colors of C, M, Y, and K in this embodiment). In this embodiment, an electrophotographic printer is used as the color image output device 130 . However, the configuration of the color image output device 130 is not limited to this, and a printer device of another system such as an inkjet system may be used.

Operation panel 140 receives an instruction input from the user and transmits it to control unit 100 . The operation panel 140 includes, for example, a display unit such as a liquid crystal display and operation keys for accepting user's operation input. Alternatively, operation panel 140 may be a touch panel.

The control unit 100 includes an arithmetic processing unit such as a CPU (Central Processing Unit) and a dedicated processor, and a storage unit such as a RAM (Random Access Memory), a ROM (Read Only Memory), and an HDD (Hard Disk Drive) (all shown in the figure). not shown). The control unit 100 receives various information stored in the storage unit, programs for executing various controls, instructions input from the user through the operation panel 140, and various sensors provided in the image forming apparatus 1. The operations of the color image input device 110, the color image processing device 120, the color image output device 130, and the operation panel 140 are controlled according to the detection result of the color image.

(Details of document feature determination unit)
The details of the document feature determination unit 13 will be described. FIG. 2 is a block diagram showing the configuration of the document feature determination section 13, and FIG. 3 is a flowchart showing the flow of processing in the document feature determination section 13. As shown in FIG.

As shown in FIG. 2, the document feature determination unit 13 includes a pixel determination unit 1301, an area pixel count unit 1302, a data selection unit 1303, a histogram generation unit 1304, a background determination unit 1305, a document type determination unit 1306, and a background removal unit. 1307 is provided.

As shown in FIG. 3, in the document feature determination unit 13, first, the pixel determination unit 1301 converts each pixel of the RGB signal (image data) input from the shading correction unit 12 into a base pixel, a photograph pixel (continuous tone pixel). ), and performs area pixel determination processing for classifying the pixels into either character pixels or halftone pixels (S1). In the following description, pixels classified into background pixels, photo pixels (continuous tone pixels), text pixels, and halftone pixels are referred to as area pixels.

Algorithms for classifying into area pixels are not particularly limited, and existing area separation methods can be used. For example, the method described in Patent Document 2 may be used. In this method, area pixel determination processing is performed by the following processing (1) to (8).

(1) Calculate the minimum density value and the maximum density value in a pixel block of n×m (for example, 7 pixels×7 pixels) including the pixel of interest.
(2) Calculate the maximum density difference using the calculated minimum density value and maximum density value.
(3) A total density complexity (for example, a sum of values calculated in the main scanning direction and the sub-scanning direction), which is the sum of the absolute values of the density differences of pixels adjacent to the target pixel, is calculated.
(4) Comparison between the calculated maximum density difference and the maximum density difference threshold and comparison between the calculated total density complexity and the total density complexity threshold are performed.
(5) When maximum density difference<maximum density difference threshold and total density complexity <sum density complexity threshold, it is determined that the pixel of interest belongs to the background/photograph area (continuous tone area).
(6) If the above conditions are not met, it is determined that the pixel of interest belongs to the character/halftone dot area.
(7) A pixel determined to belong to a background/photographic area (continuous tone area) is determined to be a background pixel when the pixel of interest satisfies the following condition: maximum density difference<background/photographic paper determination threshold. is not satisfied, it is determined to be a photographic pixel (continuous tone pixel).
(8) For pixels determined to belong to a character/halftone dot area, if the pixel of interest satisfies the condition: total density complexity < (maximum density difference multiplied by character/halftone determination threshold value), character pixel If this condition is not satisfied, it is determined to be a halftone dot pixel.
Also, in the area pixel determination process, a pixel classification process for determining the halftone ruling of the document is also performed. Details will be described later.

Note that the area pixel determination process may be performed using, for example, pre-scanned image data, or may be performed using image data temporarily stored in a storage means such as a hard disk.

Next, based on the result of the area pixel determination processing in the pixel determination unit 1301, the area pixel count unit 1302 counts the area pixels for each type (background pixel, photo pixel (continuous tone pixel), character pixel, and halftone pixel). For each area pixel of point pixels, an area pixel count process is performed to count the number of classified pixels (area pixels). The number of halftone dot ruling classification results is also counted (S2).

Next, the data selection unit 1303 separates the image area based on the RGB signals input from the shading correction unit 12 into a target area to be subjected to histogram generation processing (to be described later) and an exclusion area to be excluded from histogram generation processing. area selection processing is performed (S3). That is, in the present embodiment, histogram generation is not performed for all pixels of the input image, but is performed by excluding some pixels.

FIG. 4 is an explanatory diagram showing a method of setting an exclusion area excluded from histogram generation targets in document image data.

The image data of the document read by the color image input device 110 is processed in the shading correction unit 12 to remove various distortions caused by the illumination system, imaging system, and imaging system of the color image input device 110. Appropriate correction is not performed for the entire area of . For example, when the color image input device 110 reads a document, an area (edge area of the document) located at the edge of the document platen of the document may result in a relatively dark signal.

Therefore, in the present embodiment, as shown in FIG. 4, the data selection unit 1303 is located at the edge of the glass surface of the document platen of the color image input device 110 when reading the image data of the document. Image data within a predetermined range is excluded from the target area for histogram generation, and other areas are set as target areas for histogram generation.

The predetermined range is not particularly limited and may be set as appropriate. may be excluded from A void area is a blank area that exists with a width of about several millimeters (for example, about 4 mm) from the edge (periphery) of a document printed or copied by a printer.

Next, the histogram generation unit 1304 determines that the RGB signals input from the shading correction unit 12 belong to an area determined to be a background pixel in the area pixel determination process and to be a target area for histogram generation in the area selection process. A pixel is extracted as a pixel of interest, and histogram generation processing (minimum value histogram generation processing) is performed (S4).

Specifically, the histogram generation unit 1304 compares the pixel of interest for each plane (each color component), calculates the minimum value between each color component of the pixel of interest, , 1 is added to the frequency of the density division to which the calculated minimum value belongs. Note that in this embodiment, the number of density divisions in the histogram is defined as the first density division and the second density division in descending order of the pixel value (density value), and the division with the largest pixel value is designated as the 32nd density division, which is 32. It is set to division. In this embodiment, since RGB signals are used, the closer the pixel value (density value) is to "0" (the smaller the density value), the higher the actual density. 255” (the higher the density value), the lower the density (lighter).

Note that the histogram is created using the minimum values between the color components by taking the density value of the darkest color component of the background (the color component with the minimum density value) as the background density, and performing the correction corresponding to that density value. This is because background removal can be performed for color components other than the darkest color component by applying the amount table to all color components and performing background removal processing.

After that, the document feature determination section 13 determines whether or not the processes of S1 to S4 have been performed for all pixels (S5). The processing of S1 to S4 is performed on the pixels.

On the other hand, if it is determined in S5 that the processing of S1 to S4 has been completed for all pixels, the background determination unit 1305 performs background determination processing to determine whether background removal should be performed (S6).

After performing background determination processing in S6, the document type determination unit 1306 performs document type determination processing (S7).

Specifically, the document type determination unit 1306 determines the number of area pixels counted by the area pixel count unit 1302 and the predetermined background area, photo area (continuous tone area), halftone dot area, and character area. The type of the entire document is determined by comparing each corresponding threshold value.

For example, if the detection accuracy is higher in the order of text, halftone dots, and photographs (continuous gradation), if the ratio of pixels in the text area is 30% or more of the total number of pixels, it is determined to be a text document, and the halftone dot area is determined. If the ratio of pixels in is 20% or more of the total number of pixels, it is judged to be a halftone dot original (printed photo original), and the ratio of pixels in the photographic area (continuous tone area) is 10% or more of the total number of pixels. In the case of , it is determined that the document is a photographic paper photo document. Also, when the ratio of the text area and the ratio of the halftone dot area are equal to or greater than the respective threshold values, it is determined that the document is a text/halftone dot document (character printed photo document). When the ratio of the character area and the ratio of the photographic area (continuous gradation area) are equal to or greater than the respective threshold values, it is determined that the document is a character/photo document (character photographic paper photo document).

Also, the number of halftone dots printed on the document is determined. The details will be described later.

Note that the document type determination unit 1306 applies the document feature determination signal indicating the document type determination result to the background removal unit 1307, the input gradation correction unit 14, the color correction unit 16, the black generation under color removal unit 17, and the spatial filter processing. 18 and the gradation reproduction processing unit 20. FIG.

Next, the background removal unit 1307 determines whether or not the document should be subjected to background removal based on the result of the document type determination by the document type determination unit 1306 (S8). That is, the background removal unit 1307 determines whether or not the document type determined by the document type determination unit 1306 is a type preset as a document type for which background removal is to be performed.

For example, as a document for background removal, a text document, a document that does not include a photo area (continuous tone area), a text/halftone dot document (character printed photo document) with a small halftone dot area ratio (for example, a halftone dot area) ratio of 40% or less of the entire document) may be set.

If it is determined in S8 that the original does not require the background removal, the original document characteristics determining section 13 terminates the process without removing the background.

On the other hand, if it is determined in S8 that the document should be background-removed, the background-removing unit 1307 performs the background-removing process (S9), and ends the process.

The method of background removal processing is not particularly limited, and an existing background removal method can be used.

For example, the method described in Patent Document 3 may be used to remove the background. In this method, when the background is determined, the correction amount table corresponding to the background density value is selected, and the background removal process is performed. When a histogram is created using the minimum values of background pixels for each plane (each color component), the background removal process can be performed by selecting a correction amount table corresponding to the minimum values for all color components.

(Pixel Classification Process for Determination of Halftone Frequency of Document in Area Pixel Determination Process)
A pixel classification process for determining the halftone ruling of the document in the area pixel determination process (S1) executed by the pixel determining unit 1301 will be described. For each color component of the RGB image, in the two directions shown in FIG. On the other hand, when the density difference between adjacent pixels exceeds the adjacent pixel density difference threshold, it is determined as a rising pixel or falling pixel, and the number of determined pixels is counted as the number of inversions. However, consecutive rising pixels and falling pixels are not counted. The adjacent pixel density difference threshold is determined based on an average value in a pixel block of n×m (for example, 5 pixels×5 pixels) including the pixel of interest. If the average value is greater than or equal to a first predetermined value (eg, 100), the adjacent pixel density difference threshold is set to TH1 (eg, 50). For example, the adjacent pixel density difference threshold is set to TH2 (eg, 33), and if the average value is less than the second predetermined value, the adjacent pixel density difference threshold is set to TH3 (eg, 15). The reason why the adjacent pixel density difference threshold for counting rising pixels or falling pixels according to the average value in the pixel block is to be different is that dark pixels and light pixels are averaged depending on whether the portion is dark or light. This is because the gradation difference is different. Since there is a tendency that the density difference between dark pixels and light pixels is smaller in darker areas than in lighter areas on average, the average value is divided into three, and adjacent pixel density difference thresholds are determined for each. ing. Although the average value is divided into three here, the number of divisions is not necessarily limited to three, and the number of divisions may be four or more.

The maximum value of the inversion number count value of each component of RGB is obtained, and if the maximum value is equal to or greater than the first inversion number threshold (for example, 30), it is the first line number pixel, less than the first inversion number threshold and the second inversion number. If it is equal to or greater than the threshold value (for example, 15), it is determined to be the second line number pixel, and if it is less than the second inversion count threshold value, it is determined to be the third line number pixel. Here, the count of the number of inversions is divided into three for determination, but the number of divisions is not necessarily limited to three, and the number of divisions may be four or more. Since this is performed to determine the number of lines, it may be performed only for pixels determined to be halftone dots.

(Details of frequency determination in document type determination processing)
Screen frequency determination for determining the screen frequency of the document in the area pixel determination processing (S7) executed by the document type determination unit 1306 will be described. The following count values and histogram frequencies are used to determine the number of halftone dots printed on the document. The number of halftone dot area pixels (scr0), the number of first line number pixels (scr1), the number of second line number pixels (scr2), and the 3-line pixel number (scr3), histogram target pixel number (bgcnt) in histogram generation unit 1304, histogram 16th segment frequency (hst16), histogram 17th segment frequency (hst17), ..., histogram 31st segment frequency (hst31) Use

A feature amount (image data feature amount) relating to the image data is calculated from the above count value and histogram frequency (S71). For example, the ratio of the number of pixels of the first line number (scr1) to the number of pixels of the halftone dot area (scr0), the ratio of the number of pixels of the second line number (scr2) to the number of pixels of the halftone area (scr0), the number of pixels of the third line number Based on the ratio of (scr3) to the number of halftone dot area pixels (scr0), a feature amount (first feature amount) indicating the frequency of grayscale change in adjacent pixels is calculated. Also, the ratio of the histogram 16th segment frequency (hst16) to the number of histogram target pixels (bgcnt), the ratio of the histogram 17th segment frequency (hst17) to the histogram target pixel count (bgcnt), . . . , the histogram 31st segment frequency (hst31) to the number of pixels to be histogrammed (bgcnt), a feature quantity (second feature quantity) indicating the degree of change in pixel density in the background pixels is calculated. Based on these feature values, halftone frequency determination processing is performed to determine whether the document has a high screen frequency or a low frequency screen (S72).

Even if the number of lines of halftone dots is constant in one sheet of document, the number of inversions of densities in the mask is not constant. In addition, paper quality, density of paper, print quality, presence or absence of show-through, etc., affect the number of times of density reversal. For this reason, the feature amount of the frequency of the number of shade inversions is increased, and the density distribution is added to the feature amount.

High screen frequency/low screen frequency may be determined by machine learning based on these feature values. For example, these feature quantities are input in advance into a mathematical formula based on the learning results of the support vector machine. The support vector machine learns a pair of these feature quantities with either a binary teacher signal indicating a high frequency or a low frequency as a set. For example, sets of feature amounts for 500 sheets of originals desired to be judged as high frequency screens are prepared, and the teacher signal of these sets is set to 1. FIG. Sets of feature amounts of 500 sheets of the manuscript to be judged as having a low screen frequency are prepared, and the teacher signals of these sets are set to 0. A set of these 1,000 manuscript features and a teacher signal are set for learning. It should be noted that whether the frequency is high or low is not necessarily determined by whether the frequency is greater than or equal to a predetermined frequency. Appears or not.

なお、ｘ以外は予め学習結果により求められているものである。ｘ_ｎは１０００組の特徴量から抽出されたＮ組のサポートベクター、ａ_ｎはＮ組のサポートベクターそれぞれに対する重み、ｔ_ｎはＮ組のサポートベクターそれぞれの教師信号、ｂはバイアスと呼ばれ決定境界の位置を表すパラメータである。ｋ（ｘ，ｘ_ｎ）は、学習時にも用いる動径基底関数カーネルというもので、以下の式で定義される。

γは、学習時にも用いるパラメータである。 Based on the learning result, the vector of the set of feature values is input, and the following formula (discrimination function) is used to determine whether the number of frequencies is high or low.

Note that the parameters other than x are obtained in advance from the learning results. x _n is the N set of support vectors extracted from 1000 sets of features, a _n is the weight for each of the N sets of support vectors, t _n is the teacher signal for each of the N sets of support vectors, and b is called the bias. A parameter that represents the position of the boundary. k(x, x _n ) is a radial basis function kernel that is also used during learning, and is defined by the following equation.

γ is a parameter also used during learning.

The output result of the calculated y(x) becomes the determination result. If y(x)≧0, the number of lines is high, and if y(x)<0, the number of lines is low.

It should be noted that the high frequency/low frequency may be determined by a machine learning method other than the support vector machine. For example, the high screen frequency/low screen frequency may be determined by inputting a feature value into a formula learned by a decision tree or neural network (deep learning).

(Details of the spatial filtering unit 18)
Details of the spatial filtering unit 18 will be described. FIG. 6 is an example of a filter applied to a halftone dot area. A different filter is applied to the halftone dot area depending on whether the screen frequency determination result is a high screen frequency or a low screen frequency. A halftone dot area determined to have a high screen frequency is subjected to a process in which the degree of emphasis is higher and the smoothness is weaker than that of a halftone dot area determined to have a low screen frequency. In other words, a halftone dot area determined to have a low halftone frequency is subjected to processing that is less emphasized and smoother than a halftone dot area determined to have a high halftone frequency. FIG. 6(a) shows the filter coefficients for the halftone dot area when the screen frequency is determined to be high, and FIG. 6(b) shows the filter coefficients for the halftone dot area when the screen frequency is determined to be low.

As a result, the image data of the document determined to have a high screen frequency can be output as an image with high sharpness without appearing in the halftone dot area, which may cause a problem in image quality. Characters misrecognized in halftone dot areas and characters on halftone dots can be output clearly and easily. The image data of the document determined to have a low screen frequency can be output without the appearance of moire, which poses an image quality problem, although the sharpness of the halftone dot area is not high.

(Example of having multiple input means)
The color image input device 110 is equipped with a document platen and a document feeder, and means for reading the document by placing the document on the platen and moving the CCD, and feeding the document while the CCD is fixed by the document feeder. It may have means for reading.

Image data of a document read by a document feeder tends to cause deviation in an image between RGB as compared with image data read by a document platen. Therefore, a discrimination function for reading on the platen and a discrimination function for reading on the document feeder are separately prepared, and the feature amount is input to the discrimination function corresponding to each input means. Determine the number of lines.

Further, when the document feeder has means for reading both sides simultaneously, the reading method may be different between the front side and the back side. The front side may be read by the CCD, and the back side may be read by a CIS (Contact Image Sensor).

The reading of the front surface by CCD and the reading of the back surface by CIS do not match in color output characteristics and sharpness. Therefore, separate identification functions for reading from the front side of the document feeder and those for reading from the rear side of the document feeder are prepared separately, and the feature values are input to the discrimination functions for each case. Determine the number of lines.

Of course, each discriminant function is the result of machine learning (e.g., support vector machine) using a teacher signal and a set of feature values obtained by reading a manuscript with each input means. do.

This makes it possible to output an appropriate halftone determination result regardless of which color image input device 110 reads the image.

(Example of having multiple input resolutions)
The input resolution may differ depending on the processing instructed by the user. For example, the user may be able to select an input with 600 dpi main scanning and 600 dpi sub-scanning that gives priority to image quality and an input with 600 dpi main scanning and 400 dpi sub-scanning that gives priority to speed.

Even for the same document, the number of times of inversion is different between the input by main scanning 600 dpi and sub-scanning 600 dpi and the input by main scanning 600 dpi and sub-scanning 400 dpi. . For this reason, the number of lines is determined by each identification function for input with 600 dpi main scanning and 600 dpi sub-scanning and for input with 600 dpi main scanning and 400 dpi sub-scanning.

Of course, each discriminant function is based on the result of learning a set of feature amounts obtained by reading a document at each input resolution and a set of teacher signals.

This makes it possible to output an appropriate line number determination result regardless of the input resolution.

(Example of having one for monochrome output and one for color output)
In the case of copy output, moiré may appear due to deviations in the period of the halftone dots of the original, the period due to the input resolution, and the period due to the dither processing of the gradation reproduction processing section 20 . Although the original is read with RGB signals, the dither processing of the gradation reproduction processing unit 20 is processed only with K (black) signals in the case of monochrome output, and is processed with CMYK (C: cyan, M: magenta) in the case of color output. , Y: yellow, K: black) signals. Therefore, if the dithering process for monochromatic output is the same as that for the K (black) signal for color output, moiré tends to appear relatively less. Therefore, the number of lines is determined using the discrimination functions for monochrome output and for color output.

Of course, each discriminant function is based on the results obtained by learning sets of feature amounts obtained by reading a document with monochrome output and color output, respectively, with a teacher signal.

As a result, it is possible to output appropriate line number determination results for each of monochrome output and color output.

(Example of having multiple gradation reproduction settings)
In the case of copy output, moiré may appear due to deviations in the period of the halftone dots of the original, the period due to the input resolution, and the period due to the dither processing of the gradation reproduction processing section 20 . Therefore, the dither processing of the gradation reproduction processing unit 20 can be selected from two types with different cycles. Different dithering processes result in different moiré appearances. Moiré is relatively likely to appear when the dither processing has a rough cycle, and moire is relatively less likely to appear when the dither processing has a finer cycle. Therefore, the number of lines is determined using each of the discriminant functions for coarse-cycle dither processing and fine-cycle dither processing.

Note that the tone reproduction setting of the tone reproduction processing unit 20 may be selectable from two types of dither processing and error diffusion processing. The error diffusion process itself has no periodicity, and moire appears only in the period of the halftone dots of the original and the period due to the input resolution. Therefore, the number of lines is determined using each of the discriminant functions for dither processing and error diffusion processing.

Of course, each discriminant function is used for coarse-cycle dither processing and for fine-cycle dither processing, or for dither processing and for error diffusion processing, respectively. It is based on the result of making the set learn. It is of course possible to have tone reproduction settings other than those described above.

As a result, when there are a plurality of tone reproduction settings, it is possible to output an appropriate line number determination result based on the selected tone reproduction setting.

(for scan output)
FIG. 7 is a block diagram showing a schematic configuration of a portion related to the scan output function when the image forming apparatus 1 of FIG. 1 has the scan output function.

As shown in FIG. 7, a color image input device 110, a color image processing device 120, an operation panel 140, and a control section 100 are provided. Further, the color image processing device 120 includes an A/D conversion section 11, a shading correction section 12, a document feature determination section 13, an input gradation correction section 14, an area separation processing section 15, a color correction section 16, a spatial filter processing section 18, and a , and an output gradation correction unit 19 .

Color image input device 110, operation panel 140, and control unit 100 are the same as operation panel 140 and control unit 100 in image forming apparatus 1 of FIG.

The A/D conversion unit 11, the shading correction unit 12, the input tone correction unit 14, and the area separation processing unit 15 in the color image processing device 120 are the A/D converters of the color image processing device 120 in the image forming device 1 in FIG. Since they are the same as the conversion unit 11, the shading correction unit 12, the input tone correction unit 14, and the segmentation processing unit 15, the description thereof will be omitted.

The document feature determination unit 13 performs the same processing as the document feature determination unit 13 of the color image processing device 120 in the image forming apparatus 1 of FIG. Make a number judgment.

In the case of scan output, the dither processing of the gradation reproduction processing unit 20 is not performed, so moire is relatively less likely to appear than in copy output. Therefore, the criterion for determining whether the number of lines is high or low may be different from that for copy output. Therefore, the number of lines is determined by a discrimination function for scan output separately from that for copy output.

Of course, the discrimination function for scan output is based on the result of learning a set of feature values for scan output together with a teacher signal.

Unlike the color correction unit 16 of the color image processing device 120 in the image forming apparatus 1 of FIG. It converts to an RGB signal according to the identification signal.

The spatial filter processor 18 is different from the spatial filter processor 18 of the color image processor 120 in the image forming apparatus 1 of FIG. In order to prevent blurring and graininess deterioration of the output image by correction, spatial filtering is performed by a digital filter preset for each type of region according to the region identification signal input from the region separation processing unit 15. conduct.

The output gradation correction unit 19 differs from the output gradation correction unit 19 of the color image processing device 120 in the image forming apparatus 1 of FIG. Tone correction processing is performed.

ａは正の数で、例えば、ａ＝１で良い。 (When the user can change the threshold for determining halftone frequency)
Input the discriminant function output value y(x) to the sigmoid function.

a is a positive number, for example, a=1.

As a result, the discriminant function through the sigmoid function is output as a value greater than 0 and less than 1. Assuming that the threshold value is thl, f(y(x))≧thl is determined to be a high frequency, and f(y(x))<thl is determined to be a low frequency.

If you want to change the discriminant function according to various conditions, by passing this sigmoid function,
Output values can be standardized to fall within the range of 0 to 1.

The operation panel 140 enables the threshold to be changed according to an instruction input from the user.

FIG. 8 is an example (display example) of the operation panel 140 for setting the threshold value for line number determination. One of five levels can be set, and the initial setting is level 3, which corresponds to a threshold thl of 0.5. Changing the settings to stages 2 and 1 makes it easier to determine a lower frequency, and changing the settings to stages 4 and 5 makes it easier to determine a higher frequency. Threshold thl=0.9 for stage 1, threshold thl=0.7 for stage 2, threshold thl=0.3 for stage 4, and threshold thl=0.1 for stage 5. You may input the numerical value of a stage with an operation key, and you may touch the position of each stage with a touch panel. Also, the threshold value may be one that can be set more finely. It may be possible to directly input numerical values from 0 to 1.

As described above, according to the present embodiment, if the number of lines is determined based only on the feature value indicating the frequency of density change between adjacent pixels, the paper quality, paper density, print quality, presence or absence of show-through, halftone dot color mixture, etc. By adding a feature value that indicates the degree of shading change to the line number determination parameter, it is possible to determine the paper quality, paper density, print quality, presence or absence of show-through, etc. The result of line number judgment is less likely to be influenced by

Further, in the present embodiment, at least one of the feature quantity indicating the frequency of grayscale change between adjacent pixels and the feature quantity indicating the degree of grayscale change is a plurality of feature quantities. As a result, when it is difficult to fully capture the features of a halftone dot with only one feature value indicating the frequency of grayscale change between adjacent pixels or the degree of grayscale change, multiple types are used. By doing so, it is possible to make a judgment with high accuracy.

Further, by calculating the feature amount indicating the frequency of density change between adjacent pixels only for pixels determined to be halftone dot area pixels, it is possible to accurately determine the number of halftone dots.

In addition, by including the feature amount indicating the density distribution in the feature amount indicating the degree of change in gradation, more accurate determination can be made.

Further, by performing spatial filtering according to the line number determination result as image processing for the image data, it is possible to output an image in which moire is less likely to be perceived while maintaining sharpness.

In addition, to determine the screen frequency, the screen frequency of the halftone dots of the document is determined based on the output value obtained by inputting the image data feature amount into a mathematical expression using the results of machine learning. It is possible to make the judgment with higher accuracy than by setting a boundary and making a judgment.

This machine learning may be a support vector machine. Support vector machines are relatively simple calculations among machine learning, so they can be processed without delaying copy output or scan output.

In addition, when a plurality of image data input means are provided, the mesh of the original is obtained by inputting the image data feature amount into a mathematical expression using the result of machine learning corresponding to the input means to which the image data is input. Determine the number of lines of a point. As a result, it is possible to perform accurate determination according to the image data input means.

Further, if the image data input means has multiple image input resolution settings, based on the output value obtained by inputting the image data feature amount into a formula using the result of machine learning corresponding to the image input resolution setting, Determine the number of halftone dots of the document. As a result, it is possible to make an accurate determination according to the image input resolution.

Also, based on the output value obtained by inputting the image data feature quantity into a mathematical expression using the results of machine learning corresponding to the output format of the image data, the number of halftone dots of the document is determined. For example, by setting the output format to copy output and scan output, it is possible to output an image in which moire is less likely to be perceived while maintaining sharpness in both copy output and scan output.

In addition, a gradation reproducing means having a plurality of gradation reproduction settings is provided, and based on the output value obtained by inputting the image data feature quantity into a mathematical expression using the result of machine learning corresponding to the gradation reproduction setting, the image of the document is obtained. Determine the screen frequency of halftone dots. As a result, even if the gradation reproduction setting changes, it is possible to output an image in which the sharpness is maintained and moiré is hardly perceived.

In addition, based on the output value obtained by inputting the image data feature amount into a formula using the result of machine learning corresponding to either monochrome output or color output, the number of halftone dots of the original is calculated. judge. This makes it possible to output an image in which moire is difficult to perceive while maintaining sharpness in both monochrome output and color output.

In addition, by allowing the user to change the threshold for determining the halftone ruling for the output value of a mathematical expression using the results of machine learning, it is easier to determine a higher ruling according to the user's preference. Or conversely, it is possible to make it easier to be determined to have a low screen frequency.

The image processing apparatus according to each aspect of the present invention may be realized by a computer. In this case, an image processing program for realizing the image processing apparatus on the computer by operating the computer as the document feature determination unit 13, and a computer-readable recording medium recording it are also included in the scope of the present invention.

The present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the claims. That is, the technical scope of the present invention also includes embodiments obtained by combining technical means appropriately modified within the scope of the claims.

(processing by software)
1 and 7, the control block (particularly, the document feature determination unit 13) of the color image processing apparatus 120 may be implemented by a logic circuit (hardware) formed in an integrated circuit (IC chip) or the like, or may be realized by a CPU. may be implemented by software using

In the latter case, the color image processing apparatus 120 includes a CPU that executes instructions of a program, which is software that implements each function, and a ROM or storage device ( These are referred to as "recording media"), a RAM for developing the above programs, and the like. The object of the present invention is achieved by a computer (or CPU) reading and executing the program from the recording medium. As the recording medium, a "non-temporary tangible medium" such as a tape, disk, card, semiconductor memory, programmable logic circuit, or the like can be used. Also, the program may be supplied to the computer via any transmission medium (communication network, broadcast wave, etc.) capable of transmitting the program. It should be noted that the present invention can also be implemented in the form of a data signal embedded in a carrier wave in which the program is embodied by electronic transmission.

This makes it possible to provide a portable recording medium recording the program code (executable program, intermediate code program, source program) for executing the process of the document feature determining section 13 . Further, by reading the program recorded on the recording medium by the image forming apparatus 1, the image reading apparatus 1b, or a program reading apparatus provided in the computer system, the document feature determination processing can be executed.

The computer system includes an image input device such as a flatbed scanner, a film scanner, and a digital camera, a computer that performs various processing such as the above image processing method by loading a predetermined program code, and a CRT that displays the processing results of the computer. It may be configured by an image display device such as a display or a liquid crystal display and a printer for outputting the processing result of the computer on paper or the like. In addition, a network card, modem, or the like may be provided as communication means for connecting to a server or the like via a network.

1 image forming apparatus 100 control unit 110 color image input device 120 color image processing device 11 A/D conversion unit 12 shading correction unit 13 document feature determination unit 1301 pixel determination unit 1302 area pixel count unit 1303 data selection unit 1304 histogram generation unit 1305 Background determination unit 1306 Document type determination unit 1307 Background removal unit 14 Input gradation correction unit 15 Area separation processing unit 16 Color correction unit 17 Black generation under color removal unit 18 Spatial filter processing unit 19 Output gradation correction unit 20 Gradation reproduction processing Part 130 Color image output device 140 Operation panel

Claims (19)

an image data input unit for scanning an original and inputting it as image data;
A first feature amount indicating the frequency of grayscale change between adjacent pixels in pixels included in the halftone dot area of the image data, and a second feature amount indicating the degree of grayscale change in pixels included in the background area of the image data. a determination unit that determines the number of halftone dots of the document based on the image data feature amount including the feature amount of
an image processing unit that performs image processing on the image data according to the determination result of the determination unit;
An image processing device comprising: 2. The image processing apparatus according to claim 1, further comprising a calculator that calculates the image data feature quantity from the image data. 3. The image processing apparatus according to claim 1, wherein at least one of said first feature amount and said second feature amount includes a plurality of feature amounts. 4. The image processing according to any one of claims 1 to 3, wherein the second feature amount is calculated for pixels included in a background area other than an edge of the document. Device. 5. The image processing apparatus according to claim 1, wherein the second feature amount includes a feature amount indicating density distribution. 6. The image processing apparatus according to claim 1, wherein the image processing unit performs spatial filtering as the image processing. 2. The judgment unit judges the number of halftone dots of the document based on an output value obtained by inputting the image data feature quantity into a formula using a result of machine learning. 7. The image processing device according to any one of 6. 8. The image processing apparatus according to claim 7, wherein said machine learning is a support vector machine. comprising a plurality of the image data input units,
The determination unit determines the quality of the document based on an output value obtained by inputting the image data feature quantity into a formula using the result of the machine learning corresponding to the image data input unit to which the image data is input. 8. The image processing apparatus according to claim 7, wherein the screen frequency of halftone dots is determined. having a plurality of image input resolution settings for the image data input unit;
The determining unit determines the number of halftone dots of the document based on the output value obtained by inputting the image data feature quantity into a formula using the result of the machine learning corresponding to the image input resolution setting. 8. The image processing apparatus according to claim 7, wherein: The determination unit determines the number of halftone dots of the document based on the output value obtained by inputting the image data feature quantity into a formula using the result of the machine learning corresponding to the output format of the image data. 8. The image processing apparatus according to claim 7, wherein the determination is made. 12. The image processing apparatus according to claim 11, wherein the output format is copy output and scan output. A gradation reproduction unit having multiple gradation reproduction settings is provided,
The determination unit determines the number of halftone dots of the document based on the output value obtained by inputting the image data feature quantity into a formula using the result of the machine learning corresponding to the tone reproduction setting. 8. The image processing apparatus according to claim 7, wherein: The determination unit determines the halftone dots of the document based on the output value obtained by inputting the image data feature amount into a formula using the result of the machine learning corresponding to either monochrome output or color output. 8. The image processing apparatus according to claim 7, wherein the number of screen lines of is determined. 8. The image processing apparatus according to claim 7, wherein the determination unit is capable of changing a threshold value for determining halftone ruling with respect to an output value of a formula using the result of the machine learning. an image processing device according to any one of claims 1 to 15;
An image forming apparatus comprising an image forming section that forms an image based on image data processed by the image processing apparatus. an image data input step of scanning a document and inputting it as image data;
A first feature amount indicating the frequency of grayscale change between adjacent pixels in pixels included in the halftone dot area of the image data, and a second feature amount indicating the degree of grayscale change in pixels included in the background area of the image data. a determination step of determining the screen frequency of the halftone dots of the document based on the image data feature amount including the feature amount of
an image processing step of performing image processing on the image data according to the determination result of the determination step;
An image processing method comprising: In the processor of the image processing device,
an image data input step of scanning a document and inputting it as image data;
A first feature amount indicating the frequency of grayscale change between adjacent pixels in pixels included in the halftone dot area of the image data, and a second feature amount indicating the degree of grayscale change in pixels included in the background area of the image data. a determination step of determining the screen frequency of the halftone dots of the document based on the image data feature amount including the feature amount of
an image processing step of performing image processing on the image data according to the determination result of the determination step;
An image processing program that executes In the processor of the image processing device,
an image data input step of scanning a document and inputting it as image data;
A first feature amount indicating the frequency of grayscale change between adjacent pixels in pixels included in the halftone dot area of the image data, and a second feature amount indicating the degree of grayscale change in pixels included in the background area of the image data. a determination step of determining the screen frequency of the halftone dots of the document based on the image data feature amount including the feature amount of
an image processing step of performing image processing on the image data according to the determination result of the determination step;
A computer-readable recording medium in which an image processing program for executing is recorded.

Images