JP2002300382A - Color image processing apparatus, color image processing method, program and recording medium - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To provide a color image processing apparatus capable of correcting show-through from only one side image when scanning a color original. A color image processing apparatus according to the present invention includes a color component for binarizing the value of each color component with respect to a digital color original image obtained by digitally inputting one side of an original color-printed on both sides. A binarizing unit 30, an edge detecting unit 40 for calculating an edge intensity to detect an edge, an off-set pixel detecting unit 50 for detecting a pixel likely to be an off-set, and the color component binarizing unit Among the pixels in which all the components binarized in 30 are equal to or larger than the threshold, the edge strength is low, or only the pixels determined to be highly likely to be show-through are shown off components. A show-through correction unit 60 that generates a corrected image is provided.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a color image processing apparatus, a color image processing method, a program for executing the functions of the color image processing apparatus, and a computer-readable recording medium on which the program is recorded. “Show-through” in which the image of the second side is mixed and input into the image of the first side in a printed document
Is to be removed.

[0002]

2. Description of the Related Art FIG. 2 is a diagram showing an example of an image due to show-through. With the spread of color image input devices such as color scanners, digital cameras, and digital color copiers,
Document images of books, magazines, catalogs, advertisements, newspapers, and the like are also commonly input in color. Since these documents are often printed on both sides of a sheet of paper, as shown in FIG. 2, a "show-through" occurs in which the image on the back side is transparent and is mixed in the image on the front side. .
Further, in the case of a color document, the background color is not uniform, and a pattern, a pattern, and a photograph are often mixed. For this reason, the problem of removing “show-through” from an input image by digital image processing has been regarded as important for improving the image quality but difficult to handle. As a show-through correction method in such digital image processing, a method has been proposed in which a two-sided image is input, a show-through part is detected by comparing two images, and the luminance and color of the detected part are corrected. ing. For example, JP-A-7-87295
Gazette, JP-A-8-265563, JP-A-9-2
No. 05544, Japanese Patent Application Laid-Open No. 9-233319, and US Pat. No. 5,832,137. In addition, a method has been proposed in which only the image of the first surface is used, the distribution of edges and brightness on the image is analyzed, and a portion having a weak edge strength and a portion having a high brightness are corrected. A method of correcting the high luminance portion, for example, Japanese Patent Application Laid-Open No. 2000-2294
In Japanese Unexamined Patent Application Publication No. H06-209, an edge is detected from an input image, show-through is removed by increasing the luminance of pixels other than the edge and the low luminance area, and a corrected image is output. In the same publication, an edge is detected from an input image, an image is divided into an edge and a low-luminance region, and the luminance is increased only in a region where the average luminance in the region is high in the divided regions. The image is removed and the corrected image is output. further,
Japanese Patent Application Laid-Open No. H11-41266 proposes a method of judging presence or absence of show-through from a difference between a luminance value at the maximum frequency and a maximum luminance value obtained from a histogram of a read image.

[0003]

However, when these inventions are applied to a color image, information such as hue and saturation (or saturation) is discarded from the color information.
Since the show-through correction processing is performed using only the luminance (brightness), the show-through correction processing is also performed for yellow characters on a white background, and the hue and saturation are different from each other, but the contrast is small. Since no edge is detected for an adjacent area, the show-through correction processing may cause loss such as blurring of character boundaries and boundaries between areas, and may rather degrade image quality. Usually the second
Since the edge strength of the part shown off from the surface is often weaker than the edge intensity of the image of the first surface, the image of the first surface and the second surface are determined using a threshold determined from the distribution of the edge intensity on the image. The edge of the show-through part from is separated. However, if the threshold value of the edge strength is lowered to keep the edge of the first surface, the show-through edge also remains. Conversely, if the threshold value is increased to remove show-through edges, the edges of the first surface will also disappear. Thus, the edge of the image of the first surface and the second
It is difficult to separate a show-through edge from a surface by a threshold value, and there is always a trade-off between retaining the edge of the first surface and removing the show-through edge from the second surface. As described above, in the related art, it is difficult to automatically separate the image of the first surface from the show-through portion from the second surface, and as a result, an image area to be subjected to the show-through correction process is automatically set. It was difficult. For this reason, problems such as degradation of the image on the first surface have been caused as a side effect of the show-through correction processing, and conversely, show-through has not been sufficiently corrected. SUMMARY OF THE INVENTION The present invention is to solve the above-mentioned problem, and when scanning a color original that is printed on both sides and has a show-through from the second side on the first side, only an image on one side of the paper is input. To automatically set the image area to be show-through corrected,
A color image processing apparatus capable of correcting show-through of a color image, a color image processing method, a program for executing functions of the color image processing apparatus, and a computer-readable recording medium storing the program. Aim.

[0004]

According to a first aspect of the present invention, there is provided a color image processing apparatus comprising: a digital color original obtained by digitally inputting one side of an original color-printed on both sides; Color component binarization means for binarizing the value of each color component of an image, edge detection means for calculating edge strength and detecting edges, and detecting pixels likely to show-through Off-set pixel detection means, and among the pixels in which all the components binarized by the color component binarization means are equal to or more than a threshold, the edge strength calculated by the edge detection means is low, or A show-through correction unit that generates an image in which a show-through component is corrected only for pixels determined to be highly likely to be reflected is provided. According to a second aspect of the present invention, in the color image processing apparatus according to the first aspect, the color component binarizing unit sets a window having a predetermined size around each pixel. The threshold value is set based on a statistic of each color component of a pixel in the window. According to a third aspect of the present invention, in the color image processing device according to the first aspect, the edge detecting means includes:
The edge strength is calculated from each component of the color of the image from which the edge is detected, and the edge is detected in consideration of the correlation between the edge strengths of the components. According to a fourth aspect of the present invention, in the color image processing device according to the first aspect, the show-through pixel detecting means converts a color signal of each pixel of the input image, and Surrounding it, a window of a predetermined size is set, and a threshold determined based on a statistic for each color component of a pixel in the window is compared with a color signal after the conversion. I do. The color image processing method according to claim 5 of the present invention binarizes the value of each color component with respect to a digital color original image obtained by digitally inputting one side of an original color-printed on both sides, Detecting an edge from the edge strength, detecting a pixel that is likely to be show-through, among pixels in which all components of the binarized color are equal to or more than a threshold, the detected edge has a low intensity. Or
An image in which the show-through component is corrected is generated only for a pixel determined to be highly likely to be show-through.

According to a sixth aspect of the present invention, in the color image processing method according to the fifth aspect, a window having a predetermined size is set around each pixel, and each color component in the window is set. A threshold value is set based on the statistical amount of the color component, and the color component is binarized.
According to a seventh aspect of the present invention, in the color image processing method according to the fifth aspect, an edge strength is calculated from each component of a color of an image for detecting an edge, and a correlation between the edge strengths of each component is calculated. The edge detection is performed in consideration of the following. According to an eighth aspect of the present invention, in the color image processing method according to the fifth aspect, a color signal of each pixel of the input image is converted, and a predetermined color is determined around each pixel after the conversion. The size of the window is set, and a threshold determined based on a statistic for each color component of the pixel in this window is compared with the color signal after the conversion to detect the pixel as a pixel having a possibility of show-through. It is characterized by the following. A program according to a ninth aspect of the present invention causes a computer to function as the color image processing apparatus according to any one of the first to fourth aspects. Further, a computer-readable recording medium which records the program according to claim 10 of the present invention is a computer-readable recording medium.
Further, the image processing apparatus is caused to function as the color image processing apparatus according to any one of claims 4 to 4.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The construction and operation of an embodiment of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a functional block diagram showing an embodiment in which the present invention is applied to a color image processing apparatus (for example, a color digital copying machine).
In the first embodiment, the color image input unit 10 and the preprocessing unit 2
0, color component binarization means 30, edge detection means 40,
The show-through pixel detection unit 50, the show-through correction unit 60, and the image output unit 70 are configured. The color image input unit 10 inputs an original from a color image input device (for example, a digital color copier, a scanner, a camera, or the like), and executes A /
D-conversion processing is performed and temporarily stored in a memory or the like. Further, image data may be received from application software of a digital camera or a color document image analysis / recognition system, or input from a file in which image data is already stored. Preprocessing means 20
Performs smoothing or the like on the input image data to remove noise, and performs preprocessing such as total amount regulation and density adjustment by engine gamma. The image output unit 70 performs a dither / error diffusion process, performs halftone processing such as performing data conversion so as to conform to the characteristics of the output device, performs a layout process to obtain output data, and outputs the output data. (For example, a display device, a printer, a storage device, etc.). The above color image input means 10, preprocessing means 2
0 and the image output means 70 use the conventional technique.

The color component binarizing means 30 converts the color components (R (Red), G (Green), B
(Blue)), a binary image representing the result of classifying the input image data into ON pixels corresponding to characters, lines, dots, and picture parts on the surface and OFF pixels other than the pixels from the input image data. Output. In this binarization process, for each pixel (i, j), first, for a value R (i, j) of the R component, a character, a line,
Points and picture parts are compared under the condition R (i, j) <Tr (i, j). If the condition is satisfied, the pixel is determined to be ON pixel.
Binarization is performed by classifying the pixels as F pixels. Here, Tr (i, j) is a threshold value. For example, an average of R components in a W × W window (W is a size parameter of a window specified in advance) centered on the pixel (i, j) μr (i,
j) and the standard deviation σr (i, j), Tr (i, j) = a · μr (i, j) + b · σr (i, j). Where a and b are parameters specified in advance, and b is a parameter specified in advance as c.
It may be determined as b = c · μr (i, j). The processing for the R component is similarly performed for the G and B components. That is, a pixel (i, j) has three conditions R (i, j) <Tr (i, j), G (i, j) <Tg (i, j), and B (i, j) <Tb
If any of (i, j) is satisfied, the pixel is classified as an ON pixel, and otherwise, it is classified as an OFF pixel. In the above description, R,
Although the processing is performed in the order of G and B, the effect is the same even if the order is changed as appropriate. As a result of the color component binarization processing, the input image data is classified into ON pixels corresponding to characters, lines, dots, picture portions, and the like, and OFF pixels other than the ON pixels. For example, as a result of performing the above-described color component binarization processing on a show-through color image as shown in FIG. 2, pixels classified as ON pixels are shown in FIG. 3 as black.

FIG. 4 is a flowchart for explaining the color component binarization processing. The first pixel in the scan order of the input and preprocessed image data is
(i, j) (step S01). A binary image F (i, j) representing the result of classification into ON pixels corresponding to characters, lines, points, picture parts, and the like of this image data and other OFF pixels
Are initialized as OFF pixels (step S02). R
With respect to the components, the average μr (i, j) and the standard deviation σr (i, j) of the R component in a W × W window centered on the pixel (i, j) are calculated (step S03). For the R component, the pixel
The binarization threshold Tr (i, j) at (i, j) is calculated (step S0
4). Tr (i, j) is determined by Tr (i, j) = a · μr (i, j) + b · σr (i, j) using a and b as parameters specified in advance. It is determined whether the pixel value R (i, j) of the R component is smaller than a threshold value Tr (i, j) (step S05). When the condition in step S05 is satisfied, the flow branches to step S12, and the binary image F (i, j) is set as an ON pixel. With respect to the G component, an average μg (i, j) and a standard deviation σg (i, j) of the G component in a W × W window centered on the pixel (i, j) are calculated (step S06). For the G component, a binarization threshold value Tg (i, j) at the pixel (i, j) is calculated (step S07). Tg
(i, j) is determined by Tg (i, j) = a · μg (i, j) + b · σg (i, j) using a and b as parameters specified in advance.

It is determined whether the pixel value G (i, j) of the G component is smaller than a threshold value Tg (i, j) (step S08). When the condition of step S08 is satisfied, the flow branches to step S12, and the binary image F (i, j) is set as an ON pixel. With respect to the B component, the average μb (i, j) and the standard deviation σb (i, j) of the B component in a W × W window centered on the pixel (i, j) are calculated (step S09). For the B component, a binarization threshold Tb (i, j) at the pixel (i, j) is calculated (step S10). Tb
(i, j) is determined by Tb (i, j) = a · μb (i, j) + b · σb (i, j), using a and b as parameters specified in advance. It is determined whether the pixel value B (i, j) of the B component is smaller than a threshold value Tb (i, j) (step S11). When the condition of step S11 is satisfied, the flow branches to step S12, and the binary image F (i, j) is set as an ON pixel. If all pixels have been processed (step S13), the binary image F obtained by the above processing is output, and the processing ends (step S14). If there is a pixel that has not been processed yet, the next pixel is set to (i, j) in the scanning order, and
02 and execute the above processing (step S1).
5). The edge detecting means 40 is a color component binarizing means 3
An edge is detected for a region classified as “OFF pixel” with 0, and the pixel is determined as “high edge intensity” or “medium edge intensity” based on a statistic calculated from the distribution of the intensity of the edge. , And "small edge strength".

The following method is used to detect edges from the color image data. (1) A method of taking the square root of the sum of squares of the edge strength calculated by a Sobel operator or the like for each color component. (2) Literature H.-C. Lee and DRCok, "Detecting boundari
es in a vector field, "IEEE Trans. Signal Processin
g, vol.39, no.5, pp.1181-1194, 1991. Vector method. Of these edge detection methods,
It is known that the method (2) is more robust against noise than the method (1). Also, (2)
In the edge detection method based on the vector method, in the combination of the edge strengths calculated for each color component,
Rather than simply taking the square root of the sum of squares, the feature is to consider the correlation between components. Specifically, u (x,
If y), v (x, y), and w (x, y) are defined as the three components of color as shown in Expression 1, the edge intensity at the pixel (x, y) is given by Expression 2.

FIG. 5 shows an edge strength image generated by the Sobel operator for the image of FIG. In FIG. 5, the black density is changed as the edge strength increases. Next, the “OFF pixel” is set by the color component binarizing unit 30.
When the edge intensities are calculated for the pixels classified as, an intensity distribution as shown in FIG. 6 is obtained. From this distribution, statistics such as average μ of intensity and standard deviation σ are obtained, and based on these statistics, pixels are classified into “high edge intensity”, “medium edge intensity”, and “low edge intensity”. Classify by type. For example, when a and b are parameters such that a <b, the edge strength of the pixel is “small edge strength” when the edge strength is smaller than μ + a · σ, and “large edge strength” when the edge strength is larger than μ + b · σ. At other times, "medium edge strength" may be set.

FIG. 7 shows the result of classification for the edge strength image shown in FIG. In FIG. 7, “high edge intensity” is displayed in black, “medium edge intensity” is displayed in gray, and “edge intensity is low” is displayed in white. The show-through pixel detection means 50 includes:
Processing is performed on all color components of all pixels to detect pixels that are highly likely to be show-through, and classify each pixel into “show-off pixels” and “pixels other than show-through”. Here, in order to detect a pixel that is likely to be show-through for the purpose of detecting a character that is show-through from the second surface, first, in order to emphasize a show-through component, color detection is performed. The signal x is converted using, for example, the following function f (x) as shown in FIG. f (x) = max ｛255 * (x−t) / (255−
t), 0} Here, the color signal takes a value from 0 to 255, and the parameter t is a value between 0 and 255. With this conversion,
As shown in FIG. 9, the high luminance component is emphasized, and the low luminance component is compressed, and the show-through portion becomes clear. Color component binarization is applied to each of the converted pixels. This processing is the same as the processing procedure shown in FIG.
Instead of being classified into "N pixels" and "OFF pixels",
For (i, j), three conditions R (i, j) <Tr (i, j), G (i, j)
If any of <Tg (i, j) and B (i, j) <Tb (i, j) is satisfied, the pixel is classified as “show-off pixel”; otherwise, it is classified as “pixel other than show-through” Should be read as follows.

When the show-through pixel detecting means 50 is applied to FIG. 2, an area to be subjected to show-through correction can be detected as shown in FIG. 11 (the area to be corrected is shown in white). The show-through correction unit 60 performs show-through correction processing when the pixel detected by the show-through pixel detection unit 50 satisfies one of the following conditions. (1) “OFF pixel” and “low edge intensity” (2) “OFF pixel” and “medium edge intensity” and “off-set pixel” Here, the off-set correction processing is performed, for example, using a method applied by the present applicant. The technique of Japanese Patent Application No. 11-354412 is used. FIG.
When show-through correction is performed on a region other than the pixel indicated by black 0, an image in which show-through is corrected as shown in FIG. 11 is obtained.

<Modification of Embodiment> In the above-described embodiment,
The processing order of the color component binarization unit 30, the edge detection unit 40, and the show-through pixel detection unit 50 is serially processed. In this modification, the result of the pre-processing means 20 is converted into the three color component binarization means 30, the edge detection means 40
And the show-through pixel detection means 50 in parallel, the edge intensity for the simultaneously input image, the binarization result of the color component for the simultaneously input image,
Using the detection result of pixels that may be show-through pixels with respect to the input image, among the pixels that are equal to or more than the threshold value for all the components of the color of the binarization result, the edge strength is low or the back intensity is low. By performing the show-through correction processing only on pixels that are highly likely to be reflected, an image in which show-through has been corrected similar to the embodiment can be obtained.

<Embodiment as Program> FIG.
FIG. 2 is a schematic diagram of hardware constituting a color image processing apparatus using the present invention. In the present embodiment, first,
Each function shown in FIG. 1 is programmed and stored in the ROM 7. Then, the CPU 5 reads out and executes the program stored in the ROM 7,
Each function of the present invention is performed. When each function is executed,
A color image input from an image input device 1 such as a scanner or a digital camera is stored in a RAM 6, read out from the RAM 6, and a show-through correction image is generated as a result of a series of processing, and a printer, a color digital copying machine, etc. Is output to the image output device 4. Further, the progress of the process and the results of the process are presented to the user through the display device 2 such as a display, and when necessary, the user inputs parameters necessary for the process from the input device 3 such as a keyboard or a touch panel. . Intermediate data created during the execution of the processing is also stored in the RAM 6 and read, corrected, and written by the CPU 5 as necessary. In the above description, the recording medium of the program is the ROM 7, but other recording media, for example, a semiconductor medium (for example, ROM, IC memory card, etc.), an optical medium (for example, DVD-ROM, M
O, MD, CD-R, etc.) and magnetic media (eg, magnetic tape, flexible disk, etc.). Each program for realizing the functions of the present invention may be distributed in the form of a medium, or each program may be stored in a storage device such as a magnetic disk, and may be downloaded or pushed over a wired or wireless communication network or broadcast wave. It is also possible to distribute by.

[0016]

As described above, according to the present invention,
Color components are binarized to extract colored characters, lines, dots, picture parts, etc., and to apply show-through correction processing to the unextracted pixels to prevent image quality deterioration and prevent show-through. Can be corrected. In addition, by detecting the edge strength, the boundary between characters and lines and the portion where the background color changes are extracted, and show-through correction processing is performed on pixels having weak edge strength to correct show-through while preventing image quality deterioration. be able to. Also, pixels with low edge strength
To classify pixels that are likely to be show-through and other things and apply show-through correction processing to pixels that are likely to be show-through, thereby correcting image show-through while preventing image quality deterioration Can be. Also, by combining the binarization of the color components, the edge strength detection, and the show-through component detection, it is possible to automatically set an area where the show-through correction processing is performed.

[Brief description of the drawings]

FIG. 1 is a functional configuration diagram of a color image processing apparatus using the present invention.

FIG. 2 is a diagram illustrating an example of a show-through color image.

FIG. 3 is a diagram showing a result of binarizing color components with respect to the color image of FIG. 2;

FIG. 4 is a flowchart showing a processing procedure of color component binarization of the present invention.

FIG. 5 is a diagram showing a calculation result of a color edge intensity for the image of FIG. 2;

FIG. 6 is a diagram illustrating an example of an edge intensity distribution for “OFF pixels” with respect to the image of FIG. 2;

FIG. 7 is a diagram showing a classification result of a color edge intensity for the image of FIG. 2;

FIG. 8 is a diagram illustrating an example of a color signal conversion function for detecting show-through components.

FIG. 9 is a diagram showing a result of conversion of color signals for detecting show-through components with respect to the image of FIG. 2;

FIG. 10 is a diagram showing a result of obtaining a show-through correction target area of the image of FIG. 2;

11 is a diagram showing the result of applying show-through correction to the image of FIG. 2 by applying the present invention.

FIG. 12 is a schematic diagram illustrating a hardware configuration of a color image processing apparatus using the present invention.

[Explanation of symbols]

1 image input device, 2 display device, 3 input device, 4
Image output equipment, 5 CPU, 6 RAM, 7 ROM,
Reference Signs List 10 color image input device, 20 preprocessing means, 30
Color component binarization means, 40 edge detection means, 50
Show-through image detection means, 60 Show-through correction means, 70 Image output means

Claims (10)

[Claims] A digital color original image obtained by digitally inputting one side of an original color-printed on both sides; color component binarizing means for binarizing the value of each color component; Edge detection means for calculating the edge by calculating, show-through pixel detection means for detecting pixels likely to be show-through, all the components binarized by the color component binarization means threshold Among the pixels described above, an image in which the show-through component is corrected is generated only for a pixel determined to have a low edge intensity calculated by the edge detection means or a high possibility of the show-through. A color image processing apparatus comprising a show-through correction unit. 2. The color image processing apparatus according to claim 1, wherein the color component binarizing means sets a window of a predetermined size around each pixel, and sets a window of a pixel in the window. A color image processing apparatus, wherein a threshold value is set based on a statistic for each color component. 3. The color image processing apparatus according to claim 1, wherein said edge detecting means calculates an edge intensity from each color component of an image for detecting an edge, and calculates a correlation between the edge intensity of each component. A color image processing apparatus for detecting an edge in consideration of a relationship. 4. The color image processing apparatus according to claim 1, wherein the show-through pixel detecting means converts a color signal of each pixel of the input image, and outputs a color signal around each pixel.
A color image, wherein a window having a predetermined size is set, and a threshold is determined based on a statistic for each color component of a pixel in the window, and the color signal after the conversion is compared. Processing equipment. 5. A digital color original image obtained by digitally inputting one side of an original color-printed on both sides, binarizing the value of each color component, detecting edges from edge strength, and showing off Detecting a pixel that is likely to be, among the pixels in which all components of the binarized color are equal to or greater than a threshold, the intensity of the detected edge is low, or
A color image processing method comprising: generating an image in which a show-through component is corrected only for a pixel determined to be highly likely to be show-through. 6. The color image processing method according to claim 5, wherein a window of a predetermined size is set around each pixel, and a threshold value is set based on a statistic of each color component in the window. A color image processing method comprising: setting and binarizing the color components. 7. The color image processing method according to claim 5, wherein an edge strength is calculated from each color component of the image of which an edge is to be detected, and a correlation between the edge strengths of each component is taken into account. A color image processing method comprising detecting an edge. 8. The color image processing method according to claim 5, wherein a color signal of each pixel of the input image is converted, and a window of a predetermined size is provided around each pixel after the conversion. Is set, and a threshold determined based on a statistic for each color component of a pixel in this window is compared with the color signal after the conversion to detect the pixel as a possible show-through. Color image processing method. 9. The computer according to claim 1, wherein the computer is a computer.
A program for functioning as the color image processing apparatus according to any one of the above. 10. A computer-readable recording medium in which a program for causing a computer to function as the color image processing apparatus according to claim 1 is recorded.