JP2009188892A - Image processing apparatus, and image processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus and image processing program which removes show-through effects from the whole image accurately so that a shape of a boundary with a different density or color is preserved. <P>SOLUTION: A binarization part 11 binarizes an input image and creates a binarization image which is a binary image showing parts of a textured color and other parts. A contraction and expansion part 12 applies expansion processing after applying contraction processing, and obtains the binarization image. A region setting part 13 sets a reference region for referring to the input image so that pixel values corresponding to a target pixel in the binarization image are contained in a large number. A color estimating part 14 sets a parameter according to the size of the reference region set by the region setting part 13, estimates the original color free from show-through effects in the target pixel using the pixel values within the reference region, and outputs it as the color of the target pixel. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

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

  Conventionally, when an image of a document with images on both sides is scanned, or when an image is scanned with the documents overlaid, not only the image on the surface to be scanned but also the image on the back and other documents are scanned through. So-called show-through has occurred. As one of the techniques for removing the show-through, it is known to remove the show-through by reading images on both sides of the document and referring to the images on both sides.

  A method of removing show-through from an image on only one side has been realized by performing so-called background removal processing. In the background removal process, gradation correction is performed on a read image in accordance with a certain fixed value or a detected background level. In this case, the process is not performed in the color area other than the background, and the show-through remains.

  As another method, for example, Patent Document 1 describes that an edge in an image is detected and color threshold processing is performed on a region other than the edge to remove show-through. Further, for example, Patent Document 2 proposes that a rectangular frame surrounding a text area or a photograph is further detected, and a process for removing show-through is performed only for pixels that are not detected as a frame. In these methods, since the edge portion and the frame region are excluded from the processing target, a show-through image remains in the edge portion and the frame region.

  As a method for removing show-through from the entire image, there is a method described in Patent Document 3. In this method, a reference area including a target pixel is set, and a color frequency value is calculated from pixel values in the reference area. Several color values that are maximal are specified from the color frequency values, and the original color that is not show-through is estimated from among them. In this method, when a portion having a different density or color exists in the reference region, the shape of the boundary having a different density or color may not be saved due to an error in color estimation.

JP 2001-169080 A JP 2002-252746 A JP 2007-104286 A

  It is an object of the present invention to provide an image processing apparatus and an image processing program that correctly remove show-through from an entire image so that boundary shapes having different densities and colors are stored.

  The invention according to claim 1 includes a binarizing unit that binarizes an input image, and a reference region that includes the target pixel and includes a large number of pixel values corresponding to the target pixel in the binarized image. And a color estimation unit for estimating an original color that is not show-through based on a pixel value in a reference region set by the region setting unit in the input image. The image processing apparatus.

  According to the second aspect of the present invention, a binarization unit that binarizes an input image and a reference region that includes the target pixel and includes pixel values corresponding to the target pixel in the binarized image are set. And a color estimation unit that estimates an original color that is not show-through based on a value of a pixel in a reference region set by the region setting unit in the input image. An image processing apparatus.

  According to a third aspect of the present invention, in the image processing apparatus according to the first or second aspect of the present invention, the color estimation unit estimates the color according to the size of the reference region set by the region setting unit. An image processing apparatus that performs processing.

  According to a fourth aspect of the present invention, in the image processing apparatus according to any one of the first to third aspects of the present invention, a contraction process is further performed on the image binarized by the binarization unit. An image processing apparatus comprising: a contraction / expansion unit that performs expansion processing after being applied, wherein the region setting unit sets the reference region using a binarized image processed by the contraction / expansion unit It is.

  The invention described in claim 5 includes a binarizing unit that binarizes an input image, and a contraction / expansion unit that performs expansion processing after performing contraction processing on the image binarized by the binarization unit. Edge binary image acquisition means for acquiring the difference between the image subjected to the contraction process by the contraction / expansion means and the image subjected to the expansion process thereafter as an end binary image, and show-through from the input image A color estimation unit that estimates a non-original color, a color estimated by the color estimation unit according to the end binary image acquired by the end binary image acquisition unit, and a color before estimation An image processing apparatus having color adjustment means for performing color adjustment in between.

  In the invention described in claim 6 of the present application, the binarization means in the image processing apparatus according to any one of claims 1 to 5 converts the background color and other colors by floating binarization processing. An image processing apparatus characterized by binarization.

  The invention according to claim 7 of the present application is characterized in that the binarization means in the image processing apparatus according to any one of claims 1 to 4 binarizes by a moving average method. It is a processing device.

  The invention according to claim 8 of the present application is a binarization unit that binarizes an input image by a moving average method, and an isolation in an image obtained by performing a contraction process on an image binarized by the binarization unit Using a contraction / expansion unit that performs an expansion process so that isolated regions to which different labels are applied after labeling are not combined and a labeled binarized image processed by the contraction / expansion unit are used. An area setting means for setting a reference area so as to include many pixels of a label corresponding to the target pixel when the target pixel is included in the binarized image and includes the target pixel; An image processing apparatus comprising color estimation means for estimating an original color not show-through based on a pixel value in a reference area set by the area setting means in the input image.

  The invention according to claim 9 of the present application is a binarizing unit that binarizes an input image by a moving average method, and an isolated image in which an image binarized by the binarizing unit is subjected to a contraction process. Using a contraction / expansion unit that performs an expansion process so that isolated regions to which different labels are applied after labeling are not combined and a labeled binarized image processed by the contraction / expansion unit are used. A region setting means for setting a reference region including a pixel of a label corresponding to the target pixel when the target pixel includes the target pixel and a label is assigned to the target pixel in the binarized image; An image processing apparatus comprising color estimation means for estimating an original color not show-through based on a value of a pixel in a reference area set by the area setting means.

  In the invention according to claim 10 of the present application, the color estimation unit in the image processing apparatus according to claim 8 or 9 of the present application performs color estimation processing according to the size of the reference region set by the region setting unit. It is an image processing apparatus characterized by performing.

  The invention according to claim 11 of the present application is a binarization unit that binarizes an input image by a moving average method, and performs an expansion process after performing a contraction process on an image binarized by the binarization unit. Contraction / expansion means to be applied; color estimation means for estimating an original color not show-through from the input image; and a color estimated by the color estimation means in accordance with a binary image processed by the contraction / expansion means; An image processing apparatus comprising color adjusting means for performing color adjustment with a color before estimation.

  The invention according to claim 12 of the present application is an image processing program that causes a computer to execute the function of the image processing apparatus according to any one of claims 1 to 11.

  According to the invention described in claims 1, 2, and 5 of the present application, the show-through can be correctly removed from the entire image so that the shape of the boundary with different density and color is preserved.

  According to the invention described in claim 3 of the present application, it is possible to estimate the original color which is not show-through more correctly than in the inventions described in claims 1 and 2 of the present application.

  According to the fourth aspect of the present invention, the show-through can be removed without being affected by characters, lines, noise, or the like.

  According to the sixth aspect of the present invention, the show-through can be removed by preserving the shape of the boundary between the background color and the other colors.

  According to the invention described in claim 7 of the present application, compared to the invention described in claims 1 to 4 of the present application, it is possible to more accurately estimate the original color that is not show-through at the boundary portion having a different density or color. .

  According to the inventions described in claims 8 and 9 of the present application, it is possible to remove the show-through by storing the shape of each boundary having different densities and colors in the background color and other portions.

  According to the invention described in claim 10 of the present application, it is possible to estimate an original color that is not show-through more correctly than in the inventions described in claims 8 and 9 of the present application.

  According to the eleventh aspect of the present invention, the show-through can be removed by preserving the shapes of the borders having different densities and colors in the background color and other portions.

  According to the invention of claim 12 of the present application, the effect of the invention of any one of claims 1 to 11 of the present application can be obtained.

  FIG. 1 is a block diagram showing a configuration example of the first embodiment of the present invention. In the figure, 11 is a binarization unit, 12 is a contraction / expansion unit, 13 is an area setting unit, and 14 is a color estimation unit. The binarization unit 11 binarizes the input image so that at least a background portion and other portions can be understood. As a binarization method, a known method such as a floating binarization process may be used.

  The contraction / expansion unit 12 performs contraction processing on the image binarized by the binarization unit 11 and then performs expansion processing to obtain a foreground binarized image. The shrinking process is a process of shrinking an area other than the background portion, and characters and line drawings having a line width equal to or smaller than a predetermined width are deleted by the shrinking process. The region other than the background portion remaining after the shrinking process is returned to the original size by the expansion process.

  The region setting unit 13 sets a reference region including a target pixel that is a processing target in the color estimation unit 14. This reference area is set so that many pixel values corresponding to the target pixel in the foreground binarized image obtained by the expansion / contraction part 12 are included. For example, when the binary values are “0” and “1”, if the target pixel is “0”, the reference region is set so that many pixels with the pixel value “0” in the foreground binarized image are included. . If the target pixel is “1”, the reference area is set so that many pixels having a pixel value “1” in the foreground binarized image are included.

  Based on the value of the pixel in the reference region set by the region setting unit 13 in the input image, the color estimation unit 14 estimates the original color that is not show-through and outputs it as the color of the target pixel. At the time of estimation, a color estimation process according to the size of the reference area set by the area setting unit 13 is performed. As the color estimation process, the method described in Patent Document 3 is used.

  FIG. 2 is a block diagram illustrating an example of the color estimation unit, and FIG. 3 is an explanatory diagram illustrating an example of the operation of the color estimation unit. In the figure, 21 is a color histogram calculation unit, 22 is a color value candidate selection unit, 23 is an estimated color determination unit, and 24 is a parameter setting unit. The color estimation unit 14 sequentially sets each pixel as a target pixel to be processed in the input image, and estimates an original color that is not revealed in the target pixel.

  The color histogram calculation unit 21 calculates a color histogram of each pixel in the reference region set by the region setting unit 13. In FIG. 3A, a portion surrounded by a thick line is an example of the reference region set by the region setting unit 13, and a histogram of the color of the pixels in this reference region is calculated. The calculated histogram is a frequency value in, for example, a three-dimensional color space as shown in FIG.

  The color value candidate selection unit 22 selects an original color candidate (color value candidate) that replaces the color of the target pixel, that is, does not show through. As a process for this, first, a histogram in a predetermined partial color space is second-order differentiated from a position in the color space corresponding to the color of the pixel of interest, and the color of the convex portion of the frequency (the color whose frequency is the local maximum) Ask for. The color of the convex portion having the obtained frequency is set as a color value candidate.

  A black circle in FIG. 3B is set to a position corresponding to the color of the pixel to be processed, and a three-dimensional area (partial color space) having the black circle as a vertex is set as an object of secondary differentiation. In FIG. 3C, only the three-dimensional region that is the target of the secondary differentiation is extracted and shown, and the black circle indicates a color whose frequency becomes a convex portion by the secondary differentiation. This color is a color value candidate.

  The estimated color determination unit 23 determines an estimated color estimated as the original color before show-through from the color value candidates selected by the color value candidate selection unit 22. Several methods can be considered as the determination method. For example, among the color value candidates, the one with the highest frequency value of the color value candidate, the one with a small color difference from the color of the pixel of interest, the one with high brightness, and the like are selected. Of course, these may be combined or selected according to other conditions. Alternatively, the estimated color may be determined by various methods, such as determining an intermediate color of a plurality of color value candidates as an estimated color, or determining by weighting from the relationship with the color of the pixel to be processed.

  The parameter setting unit 24 sets parameters for processing in each of the color histogram calculation unit 21, the color value candidate selection unit 22, and the estimated color determination unit 23. Here, parameters corresponding to the size of the reference area set by the area setting unit 13 are set, and each unit performs processing according to the size of the reference area. Of course, parameters input from an input device such as a keyboard may be set, or parameters stored in advance may be set. Moreover, you may set a parameter combining these. As the parameters to be set, for example, the color value candidate selection unit 22 sets a partial color space or a threshold when determining color value candidates, and the estimated color determination unit 23 weights when determining an estimated color from the color value candidates. There is.

  FIG. 4 is a flowchart showing an example of operation in the configuration example of the first exemplary embodiment of the present invention. In S <b> 101, the binarization unit 11 binarizes the input image, and generates a binarized image that is a binary image indicating a background color portion and other portions.

  In S102, the contraction / expansion unit 12 performs a contraction process on the binarized image generated by the binarization unit 11, and then performs an expansion process to obtain a foreground binarized image.

  In S <b> 103, the region setting unit 13 sets a reference region including the target pixel in the color estimation unit 14. As described above, the reference region for referring to the input image is set so that many pixel values corresponding to the target pixel in the foreground binarized image are included. This reference area is set for each target pixel.

  In S104, the color estimation unit 14 sets a parameter according to the size of the reference region set by the region setting unit 13, and uses the pixel value in the reference region to determine the original color that is not revealed in the target pixel. Estimated and output as the color of the target pixel. This color estimation process is performed for each target pixel.

  FIG. 5 is an explanatory diagram of a specific example of the operation in the configuration example of the first exemplary embodiment of the present invention. An example of the above operation will be described using a specific example. FIG. 5A shows an example of an input image. Here, an oval figure is drawn with a certain color. For convenience of illustration, it is shown with diagonal lines. It is assumed that the background color is the paper color. As shown in this ellipse, the characters “HEA” are included in a mirror image by show-through. For the convenience of illustration, these characters are indicated by dotted lines.

  In step S101, the binarization unit 11 performs binarization processing on such an input image. In this example, binarization is performed by floating binarization processing. Thereby, the binarized image shown in FIG. 5B is obtained. Here, the ellipse part and the part of the character that show through are extracted as a part other than the background color. For convenience of illustration, these portions are shown in black.

  In S102, the contraction / expansion unit 12 performs contraction processing and expansion processing on the binarized image shown in FIG. Since the line having a predetermined line width or less disappears during the contraction process, the character portion in the binarized image is deleted when the expansion process is performed to restore the original size. Thereby, the foreground binarized image shown in FIG. 5C is obtained.

  In S103, the region setting unit 13 sets a reference region including the target pixel. FIG. 5D shows a case where the target pixel is a pixel in a background color region, and FIG. 5E shows a case where the target pixel is a pixel in a region other than the background color. In either case, a rectangular (here square) region centered on the pixel of interest is used as a basic reference region, and this rectangular region is expanded according to the value of the foreground binarized image corresponding to the pixel of interest.

  For example, in the example shown in FIG. 5D, since the target pixel is a background color pixel, the rectangular region is expanded so that many background color pixels are included. In this example, since the area a and the area b are composed of background color pixels, the rectangular area is extended to these areas and a reference area as shown in FIG. 5F is set. Note that the region c in FIG. 5D may also be included in the reference region because there are a small number of pixels other than the background color.

  In the example shown in FIG. 5E, since the target pixel is a pixel in a region other than the background color, the rectangular region is expanded so that many pixels other than the background color are included. In this example, since at least the area b is composed of background color pixels, the rectangular area is extended to these areas, and a reference area as shown in FIG. 5G is set. Note that the region a in FIG. 5E may also be included in the reference region, and a reference region as shown in FIG. 5H may be set.

  Whether each area arranged around the rectangular area is included in the reference area may be determined by comparing the number of pixels of any value with a preset threshold value. The threshold value may be different between the case where the value of the foreground binarized image of the target pixel indicates the background color and the other case. Note that the size and shape of the area to be expanded may be set arbitrarily. For example, expansion in an oblique direction may be performed.

  Using the reference region set by the region setting unit 13 in this way, in S104, the color estimation unit 14 performs parameter setting and estimation of the original color that is not revealed. The color estimation unit 14 performs color estimation using many background color pixels in the background color region, and performs color estimation using many pixels other than the background color in regions other than the background color. Accordingly, the estimation error is reduced even in the area where the background color and the other colors are in contact, and the boundary shape is preserved.

  FIG. 6 is an explanatory diagram of another specific example of the operation in the configuration example of the first exemplary embodiment of the present invention. In the specific example of the above-described operation, the case where the floating binarization process is performed as the binarization process in the binarization unit 11 is shown. Here, a specific example is shown of the case where the moving average method is used as the binarization processing in the binarization unit 11. In the moving average method, an area of a predetermined size including a pixel to be binarized is set, and a threshold is set based on an average value of the pixels in the area, and two pixels to be processed are processed. The value is to be converted. Accordingly, a boundary region having a density and color different from those of the surrounding pixels is extracted. The width of the boundary region is adjusted according to the size of the region for calculating an average value serving as a threshold value. Here, it is assumed that the binarization region is set so that a boundary region having a width wider than the reference region width set by the region setting unit 13 is obtained.

  FIG. 6A shows an example of an input image. Here, a rectangle is drawn with a certain color. For convenience of illustration, the color of the rectangular portion is indicated by diagonal lines. It is assumed that the background color is the paper color. As shown in this rectangle, the characters “HEA” are included in a mirror image by show-through. For the convenience of illustration, these characters are indicated by dotted lines.

  In step S101, the binarization unit 11 performs binarization processing on the input image using the moving average method. Thereby, the binarized image shown in FIG. 6B is obtained. Here, an area having a predetermined width is extracted at a rectangular boundary. For convenience of illustration, these portions are shown in black. In this example, the show-through character portion is also extracted as having a different density and color from the surroundings.

  In S102, the contraction / expansion unit 12 performs contraction processing and expansion processing on the binarized image shown in FIG. Since the line having a predetermined line width or less disappears during the contraction process, the character portion in the binarized image is deleted when the expansion process is performed to restore the original size. Thereby, the binarized image shown in FIG. 6C is obtained.

  In S103, the region setting unit 13 sets a reference region including the target pixel. FIG. 6D shows a case where the target pixel is a pixel in the background color area, and FIG. 6F shows a case where the target pixel is a pixel inside the area other than the background color. E) shows a case where the target pixel is a pixel in the boundary region obtained by the binarization process. In either case, a rectangular (here square) region centered on the target pixel is used as a basic reference region, and this rectangular region is expanded according to the value of the binarized image corresponding to the target pixel.

  For example, in the example shown in FIG. 6D, since the target pixel is a pixel in the background color area, the rectangular area is expanded so that many background color pixels are included. In this example, since the areas a and b are composed of background color pixels, the rectangular area is extended to these areas and a reference area as shown in FIG. 6G is set. Also, in the example shown in FIG. 6F, the rectangular area is expanded so that many pixels of colors other than the background color are included. In this example, the rectangular area is expanded to areas a and b, and a reference area as shown in FIG. 6I is set. In FIG. 6D, the region c may be included in the reference region.

  In the example shown in FIG. 6E, the target pixel is a pixel in the boundary region. In this case, the rectangular area is expanded so that many pixels in the boundary area are included. In this example, since the areas a and c are composed of pixels in the boundary area, the rectangular area is extended to these areas and a reference area as shown in FIG. 6H is set.

  Whether each area arranged around the rectangular area is included in the reference area may be determined by comparing the number of pixels of any value with a preset threshold value. The threshold value may be varied depending on whether or not the target pixel is a boundary region obtained by the binarization process. Note that the size and shape of the area to be expanded may be set arbitrarily. For example, expansion in an oblique direction may be performed.

  Using the reference region set by the region setting unit 13 in this way, in S104, the color estimation unit 14 performs parameter setting and estimation of the original color that is not revealed. The color estimation unit 14 uses many background color pixels in the background color region, and performs color estimation using many pixels other than the background color in the region other than the background color. Further, color estimation is performed using a large number of pixels in the boundary area. Therefore, the estimation error in the boundary region is reduced, the original color is accurately estimated, and the boundary shape is preserved.

  FIG. 7 is a flowchart showing another operation example in the configuration example of the first exemplary embodiment of the present invention. As described above, the boundary region is extracted by performing the binarization process by the moving average method, and this boundary region is also extracted for the boundary portion having a different color existing in the region other than the background color. In this operation example, an example is shown in which the accuracy of color estimation is improved for a boundary portion existing in a region other than the background color.

  In S <b> 111, the binarization unit 11 binarizes the input image by a moving average method, and generates a binarized image that is a binary image indicating a boundary portion where the density and color change.

  In S112, the contraction / expansion unit 12 performs contraction processing on the binarized image generated by the binarization unit 11, and labels an isolated region. The labeling method is arbitrary, and a known method may be applied.

  Further, in S113, the contraction / expansion unit 12 performs an expansion process so that isolated regions to which different labels are attached are not combined, thereby obtaining a foreground binarized image.

  In S <b> 114, the region setting unit 13 sets a reference region including the target pixel in the color estimation unit 14. The reference area is set such that when a pixel of interest in the foreground binarized image is labeled, a number of pixels with a label corresponding to the pixel of interest are included. In addition, when the label is not provided to the target pixel, the reference area may not be expanded, or the reference area may be set so that many pixels not provided with the label are included. The reference area is set for each target pixel.

  In S115, the color estimation unit 14 sets a parameter according to the size of the reference region set by the region setting unit 13, and uses the pixel value in the reference region to determine the original color that is not revealed in the target pixel. Estimated and output as the color of the pixel of interest. This color estimation process is performed for each target pixel.

  FIG. 8 is an explanatory diagram of still another specific example of the operation in the configuration example of the first exemplary embodiment of the present invention. FIG. 8A shows an example of an input image. Here, an example is shown in which a rectangle is drawn with a certain color, and a rectangle is drawn with a different color at a part thereof. For convenience of illustration, the color of each rectangular portion is indicated by diagonal lines. It is assumed that the background color is the paper color. As shown in these rectangles, the characters “HEA” are included in the mirror image by show-through. For the convenience of illustration, these characters are indicated by dotted lines. In this example, it is assumed that a boundary region having a width narrower than the width of the reference region having a basic size set by the region setting unit 13 is obtained by binarization processing using the moving average method.

  In step S111, the binarization unit 11 performs binarization processing on the input image by the moving average method. Thereby, the binarized image shown in FIG. 8B is obtained. Here, a region having a predetermined width is extracted at each rectangular boundary. For convenience of illustration, these portions are shown in black. In this example, the show-through character portion is also extracted as having a different density and color from the surroundings.

  In S <b> 112, the contraction / expansion unit 12 performs contraction processing on the binarized image illustrated in FIG. 8B to label isolated regions. In the shrinking process, the line having a predetermined line width or less disappears, so that the character portion is deleted and a rectangular boundary portion remains. A labeling process is performed on the boundary portion of the rectangle. By this labeling process, each rectangular boundary region is distinguished. The binarized image shown in FIG. 8C is obtained by the contraction process, and labels are attached to the isolated regions. Here, the label 1 is attached to the boundary portion of the small rectangle, and the label 2 is attached to the boundary portion of the large rectangular background.

  In step S113, the contraction / expansion unit 12 performs an expansion process on the binarized image that has undergone the contraction process. Thereby, the binarized image shown in FIG. 8D is obtained. Note that the label is also extended and attached to the expanded portion.

  In S114, the region setting unit 13 sets a reference region including the target pixel. FIG. 8E shows a case where the target pixel is a pixel in the boundary region. As a basic reference area, a rectangular area (in this case, a square) centered on the target pixel is used, and the reference area is expanded in this rectangular area so that many pixels with a label corresponding to the target pixel are included.

  For example, in the example shown in FIG. 8E, since the label corresponding to the target pixel is label 1, the basic reference area is expanded so that many pixels of label 1 are included. In this example, since many pixels of label 1 are included in areas a and c, the rectangular area is expanded to these areas, and a reference area as shown in FIG. 8F is set. In addition, although the pixel of the boundary area | region is contained about the area | region b, since it is a different label (label 2), the expansion to this area | region is not performed.

  When the target pixel is not labeled, the basic reference area (rectangular area) may be applied as it is, or the reference area is expanded to include many pixels that are not labeled. May be.

  Whether or not to include each area arranged around the rectangular area in the reference area is based on the number of pixels to which the label corresponding to the target pixel is assigned (or no label is assigned) and a preset threshold value. What is necessary is just to compare. The threshold value may be varied depending on whether or not the target pixel is a boundary region obtained by the binarization process. Note that the size and shape of the area to be expanded may be set arbitrarily. For example, expansion in an oblique direction may be performed.

  Using the reference region set by the region setting unit 13 in this way, in S115, the color estimation unit 14 performs parameter setting and estimation of the original color that is not revealed. The color estimation unit 14 uses many background color pixels in the background color region, and performs color estimation using many pixels other than the background color in the region other than the background color. Further, for the boundary region, color estimation is performed by using many pixels in the boundary portion whose density and color are similar to the target pixel. Therefore, the estimation error in the boundary region is reduced, the original color is accurately estimated, and the boundary shape is preserved.

  Next, a second embodiment of the present invention will be described. In the second embodiment, the reference region setting method in the region setting unit 13 is different from the configuration of the first embodiment shown in FIG. The area setting unit 13 in the second embodiment sets a reference area composed of pixel values corresponding to the target pixel in the foreground binarized image obtained by the expansion / contraction unit 12. For example, when the binary values are “0” and “1”, if the pixel of interest is “0”, a reference region composed of pixels having a pixel value of “0” in the foreground binarized image is set. If the target pixel is “1”, a reference area composed of pixels having a pixel value of “1” in the foreground binarized image is set.

  As an operation example of the second embodiment, as the processing of S103 in the flowchart of the first embodiment shown in FIG. 4, the area setting unit 13 uses a pixel corresponding to the target pixel in the foreground binarized image. A reference area composed of values may be set.

  FIG. 9 is an explanatory diagram of a specific example of the operation in the configuration example of the second exemplary embodiment of the present invention. As described in the first embodiment, when the image shown in FIG. 5A is input, the binarization unit 11 performs binarization processing and the binary shown in FIG. 5B. A converted image is obtained. Here, a binarized image is obtained by floating binarization. Further, the contraction / expansion unit 12 performs contraction processing and expansion processing on the binarized image shown in FIG. 5B, and the foreground binarized image shown in FIG. 5C is obtained.

  Based on the obtained foreground binarized image, the region setting unit 13 sets a reference region including the target pixel. FIG. 9A shows a basic reference region centered on the target pixel when the target pixel is a pixel in the background color region and when the target pixel is a pixel in a region other than the background color. Yes. In either case, a rectangular (here, square) region centered on the target pixel is used as a basic reference region. Of the pixels in this basic reference area, an area composed of pixels having pixel values corresponding to the target pixel is set as a reference area.

  For example, the example shown in FIG. 9C shows a case where the target pixel is a pixel in the background color area. In the basic reference region when the pixel of interest shown in FIG. 9A is a background color, a region other than the background color (a region painted black in the drawing) is included in part. An area other than the background color may be removed, and a portion composed of background color pixels as shown in FIG. 9C may be set as a reference area.

  Further, the example shown in FIG. 9B shows a case where the target pixel is a pixel in a region other than the background color. In the basic reference region when the pixel of interest shown in FIG. 9A is other than the background color, a background color region (white region in the drawing) is partially included. An area other than the background color may be removed, and a portion composed of pixels other than the background color as shown in FIG. 9B may be set as a reference area.

  Using the reference region set in this way, the color estimation unit 14 may perform parameter setting and estimation of the original color that is not revealed. The color estimation unit 14 performs color estimation using a background color pixel in the background color region, and performs color estimation using a pixel other than the background color for a region other than the background color. Accordingly, the estimation error is reduced even in the area where the background color and the other colors are in contact, and the boundary shape is preserved.

  FIG. 10 is an explanatory diagram of another specific example of the operation in the configuration example of the second exemplary embodiment of the present invention. In this specific example, the binarization unit 11 performs binarization processing by the moving average method, and the width of the boundary region extracted by binarization is the basic reference region as described in FIG. It is assumed to be wider than the size of. The image binarized by the binarization unit 11 is further subjected to reduction processing and expansion processing by the contraction / expansion unit 12 to obtain a foreground binarized image shown in FIG.

  Based on the obtained foreground binarized image, the region setting unit 13 sets a reference region including the target pixel. FIG. 10A shows a case where the target pixel is a pixel in the background color region, a case where the target pixel is a pixel in the boundary region among regions other than the background color, and a case other than that. The basic reference area is shown. In either case, a rectangular (here, square) region centered on the target pixel is used as a basic reference region. Of the pixels in this basic reference area, an area composed of pixels having pixel values corresponding to the target pixel is set as a reference area.

  For example, the example illustrated in FIG. 10B illustrates a case where the target pixel is a pixel in the background color area. In the basic reference region when the pixel of interest shown in FIG. 10A is the background color, a region other than the background color (a border region filled in black in the drawing) is partially included. An area other than the background color may be removed, and a portion composed of background color pixels as shown in FIG. 10B may be set as a reference area. In addition, the example illustrated in FIG. 10D illustrates a case where the target pixel is an internal pixel other than the boundary region among regions other than the background color. In this case, the boundary area filled in black in the drawing is partially included. An area other than the background color may be removed, and a part constituted by pixels outside the boundary area in the area other than the background color as shown in FIG. 10D may be set as the reference area.

  Furthermore, the example shown in FIG. 10C shows a case where the target pixel is a pixel in a boundary region other than the background color. In this example, the target pixel is a pixel other than the background color, but is treated separately from the internal pixel. Therefore, when the target pixel is a pixel in the boundary area, a reference area as shown in FIG. 10C is set so that pixels other than the boundary area are not included even if the area is other than the background color. To do.

  Using the reference region set in this way, the color estimation unit 14 performs parameter setting and estimation of the original color that is not revealed. In the background color region, the color estimation unit 14 performs color estimation using the background color pixels, and among the regions other than the background color, color estimation is performed using the pixels in the region other than the boundary region. Will do. Further, color estimation is performed for the boundary region using pixels in the boundary region. Therefore, color estimation is performed using pixels having similar characteristics in each region, and there is little estimation error, and the boundary shape is preserved.

  FIG. 11 is an explanatory diagram of still another specific example of the operation in the configuration example of the second exemplary embodiment of the present invention. This specific example also shows an example in which the binarization unit 11 performs binarization processing by the moving average method, and the size of the basic reference area is binarized as in the example shown in FIG. The case where it is larger than the width | variety of the boundary area | region extracted by FIG.

  The image binarized by the binarization unit 11 shown in FIG. 8 (A) is subjected to a reduction process by the contraction / expansion unit 12 and then a labeling process is performed on the isolated region, and the expansion process is further performed. A foreground binarized image shown in FIG. 11A is obtained. This is the same as FIG.

  Based on the obtained foreground binarized image, the region setting unit 13 sets a reference region including the target pixel. FIG. 11A shows a case where the target pixel is a pixel in the boundary area. As a basic reference area, a rectangular (here, square) area centered on the target pixel is used. Of the pixels in this basic reference area, an area composed of pixels assigned labels corresponding to the target pixel is set as a reference area.

  For example, the example shown in FIG. 11A shows a case where the label corresponding to the target pixel is label 1, and a reference area composed of pixels of label 1 is set out of the basic reference area pixels. To do. In addition, even if it is a pixel of a border region, the pixel to which the other label is provided is excluded. As a result, a reference area as shown in FIG. 11B is set.

  Although not shown, when the target pixel is not labeled, the basic reference region (rectangular region) may be applied as it is, or the pixel to which the label is added is excluded. You may set the reference area | region comprised by the pixel to which the label is not provided.

  Using the reference region set in this way, the color estimation unit 14 performs parameter setting and estimation of the original color that is not revealed. In the background color region, the color estimation unit 14 performs color estimation using the background color pixels, and among the regions other than the background color, color estimation is performed using the pixels in the region other than the boundary region. Will do. Furthermore, because the reference area is set according to the result of labeling for the boundary area, color estimation is performed using pixels with similar colors, and there is little estimation error, and the boundary shape is saved. Will be.

  Note that, in any of the binarization methods, if there are few pixels included in the set reference area, an error occurs when the original color that is not show-through is estimated by the color estimation unit 14. There is. For example, the number of samples when the color histogram calculation unit 21 calculates the color histogram is reduced, and color value candidates must be extracted from a small frequency value. For this reason, the extracted color value candidates vary, and an error is likely to occur when an estimated color is determined from the color value candidates. In order to deal with such a case, for example, as described in the first embodiment described above, a method of extending the reference region so that the number of pixels corresponding to the target pixel increases may be used in combination.

  FIG. 12 is a block diagram illustrating a configuration example of the third exemplary embodiment of the present invention. In the figure, 31 is a binarization unit, 32 is a contraction processing unit, 33 is an expansion processing unit, 34 is an end binary image acquisition unit, 35 is a color estimation unit, and 36 is a color adjustment unit. The binarization unit 31 binarizes the input image so that at least the background portion and other portions can be understood. As a binarization method, a known method such as a floating binarization process may be used.

  The contraction processing unit 32 performs contraction processing on the image binarized by the binarization unit 31. By this contraction process, the region other than the background portion is contracted. With this shrinking process, characters and line drawings whose line width is equal to or smaller than a predetermined width are deleted.

  The expansion processing unit 33 performs expansion processing on the binarized image that has been subjected to the contraction processing by the contraction processing unit 32. The region other than the background portion remaining after the contraction processing by the contraction processing unit 32 is returned to the original size region by the expansion processing.

  The end binary image acquisition unit 34 calculates the difference between the binarized image subjected to the contraction process by the contraction processing unit 32 and the binarized image subjected to the expansion process by the expansion processing unit 33 thereafter to the end binary. Obtain as an image. The difference between the binarized image that has been subjected to the contraction process and the binarized image that has been subjected to the expansion process thereafter indicates an end region other than the background portion.

  The color estimation unit 35 uses each pixel of the input image as a target pixel, and estimates an original color that is not show-through based on the value of a pixel in a predetermined reference area including the target pixel. As the color estimation unit 35, the configuration shown in FIG.

  The color adjustment unit 36 performs color adjustment between the color estimated by the color estimation unit 35 and the color before estimation according to the end binary image acquired by the end binary image acquisition unit 34. As described above, the edge binary image shows an edge area of an area other than the background part, and in that edge area, the estimated color is not output as it is, but a color close to the color before estimation is output. It may be adjusted as follows.

  In this third embodiment, the estimated color obtained by the color estimation unit 35 is determined uniformly by setting the reference region even in the boundary portion between the background color and the other regions. An estimation error may occur, and the contour shape may collapse. By adjusting the color adjusting unit 36, the collapse of the contour shape in the boundary region is made inconspicuous.

  FIG. 13 is a flowchart showing an example of operation in the configuration example of the third exemplary embodiment of the present invention. In S121, the binarization unit 31 binarizes the input image, and generates a binarized image that is a binary image indicating the background color portion and the other portions.

  In S122, the contraction processing unit 32 performs contraction processing on the binarized image generated by the binarization unit 31, and then in S123, the expansion processing unit 33 performs expansion processing. Further, in S124, the end binary image acquisition unit 34 calculates a difference between the binarized image subjected to the reduction process in S122 and the binarized image further subjected to the expansion process in S123 to obtain an end binary image. To get. In this edge binary image, an edge area out of areas other than the background color is extracted.

  The processing from here is performed using each pixel of the input image as a target pixel. In S125, the color estimation unit 35 estimates an original color that is not show-through in the target pixel. In S126, the color adjusting unit 36 performs color adjustment between the color estimated in S125 and the color before estimation based on the edge binary image acquired in S124. More specifically, with reference to the edge binary image acquired in S124, and the pixel of the edge binary image corresponding to the target pixel indicates a region other than the edge region, the estimated color obtained in S125 Is output. When the pixel of the end binary image corresponding to the target pixel indicates the end region, the color is adjusted to a color between the estimated color and the color before estimation. The adjustment amount may be set arbitrarily. By this adjustment, an error in color estimation generated at the end of the region outside the end region is reduced and made inconspicuous.

  FIG. 14 is an explanatory diagram of a specific example of the operation in the configuration example of the third exemplary embodiment of the present invention. FIG. 14A shows an example of the image shown in FIG. 5A, and it is assumed that this image has been input. For convenience of illustration, the color of the ellipse figure is indicated by hatching, and the show-through character is indicated by a dotted line.

  The binarization unit 31 performs binarization processing on such an input image in S121, thereby obtaining a binarized image shown in FIG. Here, the ellipse part and the part of the character that show through are extracted as a part other than the background color. For convenience of illustration, these portions are shown in black.

  In S122, the contraction processing unit 32 performs contraction processing on the binarized image shown in FIG. The result is shown in FIG. During the shrinking process, the line below the predetermined line width disappears. In addition, the oval portion is reduced by the shrinking process.

  In step S123, the expansion processing unit 33 performs expansion processing on the image on which the contraction processing illustrated in FIG. The result is shown in FIG. By this expansion process, the elliptical portion returns to its original size. Note that the character portion disappears during the contraction process, and therefore is not reproduced even if the expansion process is performed.

  In S124, an end binary image is acquired from the difference between the binarized image subjected to the contraction process and the binarized image subjected to the expansion process. As described above, in the binarized image subjected to the shrinking process, the oval part is shrunk and becomes smaller than the original size, so the difference from the binarized image that was restored to the original size by the dilation process That is, a portion that is within the ellipse in the binarized image after the expansion process but is outside the ellipse is extracted from the binarized image after the contraction process. Thereby, the area | region of the edge part of an ellipse is extracted as shown in FIG.14 (E).

  When the end binary image is obtained in this way, the color before show-through is estimated in S125 with each pixel of the input image as the target pixel. Then, color adjustment is performed in S126 for the obtained estimated color. When the pixel of interest is an edge region other than the background color, that is, a pixel in the region shown in black in the edge binary image shown in FIG. 14E, between the estimated color and the original color Perform color adjustment. In other cases, the estimated color is output.

  As described above, when the color estimation unit 35 sets a predetermined reference area with the configuration shown in FIG. 2 and estimates the original color without show-through, the reference is made at the boundary between the background color area and the other areas. An estimation error may occur because both pixels are mixed in a region. As described in the third embodiment, an estimation error occurs in the end region other than the background color by performing color adjustment between the estimated color and the original color. Is also inconspicuous, preventing the shape of the boundary from collapsing.

  FIG. 15 is a block diagram showing a modification of the third embodiment of the present invention. In the modification of the third embodiment, the binarization unit 41 binarizes the input image by the moving average method to obtain a binarized image. As already described, when binarization processing is performed using the moving average method, a boundary region is extracted as a binarized image. Therefore, the end binary image acquisition unit 34 is not provided. Accordingly, the color adjustment unit 36 uses the binarized image after the expansion processing by the expansion processing unit 33 as an end binary image, and is estimated by the color estimation unit 35 according to the end binary image. Color adjustment is performed between the color and the color before estimation.

  FIG. 16 is a flowchart showing an example of the operation in the configuration example of the modified example of the third embodiment of the present invention. In S131, the binarization unit 41 binarizes the input image by a moving average method, and generates a binarized image that is a binary image indicating a boundary portion in which density and color change.

  In S <b> 132, the contraction processing unit 32 performs contraction processing on the binarized image generated by the binarization unit 41, and then the expansion processing unit 33 performs expansion processing. Let the obtained binarized image be an end binary image.

  On the other hand, in S133, the color estimation unit 35 estimates an original color that is not show-through in the pixel of interest and sets it as an estimated color.

  In S134, the color adjustment unit 36 performs color adjustment between the estimated color obtained in S133 and the color before estimation based on the end binary image after the contraction / expansion processing in S132. More specifically, referring to the end binary image, if the value corresponding to the target pixel indicates a region other than the density or color boundary region, the estimated color obtained in S133 is output. If the value corresponding to the pixel of interest indicates a density or color boundary region, the color is adjusted to a color between the estimated color and the color before estimation. The adjustment amount may be set arbitrarily. By this adjustment, an error in color estimation generated in the boundary region where the density and color change is reduced and made inconspicuous.

  FIG. 17 is an explanatory diagram of a specific example of the operation in the configuration example of the modified example of the third embodiment of the present invention. FIG. 17A shows an example of an input image. Here, a rectangle is drawn with a certain color, and further, a rectangle is drawn with a different color. For the convenience of illustration, the diagonal lines are different to show the color difference. It is assumed that the background color is the paper color. As shown in these rectangles, the characters “HEA” are included in the mirror image by show-through. For the convenience of illustration, these characters are indicated by dotted lines.

  In step S131, the binarization unit 41 performs binarization processing on the input image using the moving average method. Thereby, the binarized image shown in FIG. Here, the boundary part and the part of the show-through character which are different in color are extracted. For convenience of illustration, these portions are shown in black.

  In S132, the contraction processing unit 32 performs contraction processing on the binarized image shown in FIG. 17B, and the expansion processing unit 33 further performs expansion processing. Since the line having a predetermined line width or less disappears during the contraction process, the character portion in the binarized image is deleted when the expansion process is performed to restore the original thickness. As a result, the end binary image shown in FIG. 17C is obtained.

  On the other hand, the color estimation unit 35 uses each pixel of the input image as a target pixel, and in S133, estimates a color before show-through to obtain an estimated color. Then, color adjustment is performed in S134 for the obtained estimated color. If the pixel of interest is a boundary region where the density or color changes, that is, a pixel in the region shown in black in the edge binary image shown in FIG. 17C, the color between the estimated color and the original color Make adjustments. In other cases, the estimated color is output.

  As described above, when the color estimation unit 35 sets a predetermined reference area with the configuration shown in FIG. 2 and estimates the original color without show-through, an estimation error occurs in the boundary area where the density and color change greatly. There is a case. Also in the modified example of the third embodiment, the boundary region is not noticeable even if an estimation error occurs by performing color adjustment between the estimated color and the original color, and the shape of the boundary portion is reduced. Prevents collapse.

  FIG. 18 is an explanatory diagram of an example of a computer program, a storage medium storing the computer program, and a computer when the functions described in the embodiments of the present invention and the modifications thereof are realized by a computer program. In the figure, 51 is a program, 52 is a computer, 61 is a magneto-optical disk, 62 is an optical disk, 63 is a magnetic disk, 64 is a memory, 71 is a CPU, 72 is an internal memory, 73 is a reading unit, 74 is a hard disk, and 75 is An interface 76 is a communication unit.

  A part or all of the functions of the units described in the above-described embodiments and modifications thereof may be realized by a program 51 that can be executed by a computer. In that case, the program 51 and data used by the program may be stored in a computer-readable storage medium. The storage medium is a signal of a corresponding signal that causes a change state of energy such as magnetism, light, electricity, etc. according to the description content of the program to the reading unit 73 provided in the hardware resource of the computer. In this format, the program description can be transmitted to the reading unit 73. For example, there are a magneto-optical disk 61, an optical disk 62 (including a CD and a DVD), a magnetic disk 63, a memory 64 (including an IC card and a memory card), and the like. Of course, these storage media are not limited to portable types.

  The program 51 is stored in these storage media, and the program 51 is read from the computer and stored in the internal memory 72 or the hard disk 74 by, for example, mounting these storage media on the reading unit 73 or the interface 75 of the computer 52. By executing the program 51 by the CPU 71, all or a part of the functions described in the above embodiments and modifications thereof may be realized. Alternatively, the program 51 is transferred to the computer 52 via a network or the like, and the computer 52 receives the program 51 by the communication unit 76 and stores the program 51 in the internal memory 72 or the hard disk 74. You may implement | achieve all or one part of the function demonstrated by each embodiment and its modification. In addition, various devices may be connected to the computer 52 via the interface 75. For example, a display device for displaying information and an input device for inputting information by the user are also connected.

  Of course, some functions may be configured by hardware, or all may be configured by hardware. Or you may comprise the program 51 with another structure. For example, an apparatus including an image reading apparatus such as a copying machine or an image forming apparatus may be configured as one program together with a control program so that show-through is removed from an image including a show-through read by the image reading apparatus. Good. Of course, when applied to other purposes, it may be integrated with the program for that purpose.

It is a block diagram which shows the structural example of the 1st Embodiment of this invention. It is a block diagram which shows an example of a color estimation part. It is explanatory drawing of an example of operation | movement of a color estimation part. It is a flowchart which shows an example of the operation | movement in the structural example of the 1st Embodiment of this invention. It is explanatory drawing of the specific example of operation | movement in the structural example of the 1st Embodiment of this invention. It is explanatory drawing of another specific example of the operation | movement in the structural example of the 1st Embodiment of this invention. It is a flowchart which shows another operation example in the structural example of the 1st Embodiment of this invention. It is explanatory drawing of another specific example of the operation | movement in the structural example of the 1st Embodiment of this invention. It is explanatory drawing of the specific example of operation | movement in the structural example of the 2nd Embodiment of this invention. It is explanatory drawing of another specific example of the operation | movement in the structural example of the 2nd Embodiment of this invention. It is explanatory drawing of another specific example of the operation | movement in the structural example of the 2nd Embodiment of this invention. It is a block diagram which shows the structural example of the 3rd Embodiment of this invention. It is a flowchart which shows an example of the operation | movement in the structural example of the 3rd Embodiment of this invention. It is explanatory drawing of the specific example of operation | movement in the structural example of the 3rd Embodiment of this invention. It is a block diagram which shows the modification of the 3rd Embodiment of this invention. It is a flowchart which shows an example of the operation | movement in the structural example of the modification of the 3rd Embodiment of this invention. It is explanatory drawing of the specific example of operation | movement in the structural example of the modification of the 3rd Embodiment of this invention. FIG. 20 is an explanatory diagram of an example of a computer program, a storage medium storing the computer program, and a computer when the functions described in the embodiments of the present invention and the modifications thereof are realized by a computer program.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 11 ... Binarization part, 12 ... Contraction expansion part, 13 ... Area setting part, 14 ... Color estimation part, 21 ... Color histogram calculation part, 22 ... Color value candidate selection part, 23 ... Estimated color determination part, 24 ... Parameter Setting unit 31... Binarization unit 32. Shrinkage processing unit 33. Expansion processing unit 34. End binary image acquisition unit 35. Color estimation unit 36. Color adjustment unit 51. Computer 61 ... Magneto-optical disk 62 ... Optical disk 63 ... Magnetic disk 64 ... Memory 71 ... CPU 72 ... Internal memory 73 ... Reading unit 74 ... Hard disk 75 ... Interface 76 ... Communication unit

Claims (12)

  Binarization means for binarizing an input image; area setting means for setting a reference area so as to include many pixel values corresponding to the target pixel in the binarized image including the target pixel; and the input An image processing apparatus comprising color estimation means for estimating an original color that is not show-through based on a value of a pixel in a reference area set by the area setting means in the image.   Binarization means for binarizing an input image, area setting means for setting a reference area including pixel values corresponding to the pixel of interest in the binarized image including the pixel of interest, and in the input image An image processing apparatus comprising color estimation means for estimating an original color that is not show-through based on a value of a pixel in a reference area set by the area setting means.   The image processing apparatus according to claim 1, wherein the color estimation unit performs a color estimation process according to a size of a reference area set by the area setting unit.   The image processing apparatus further includes a contraction / expansion unit that performs an expansion process after performing a contraction process on the image binarized by the binarization unit, and the region setting unit performs binarization after the process by the contraction / expansion unit The image processing apparatus according to claim 1, wherein the reference area is set using an image.   A binarization unit that binarizes an input image, a contraction / expansion unit that performs an expansion process on the image binarized by the binarization unit, and a contraction process performed by the contraction / expansion unit End binary image acquisition means for acquiring a difference between the applied image and an image subjected to expansion processing thereafter as an end binary image, and color estimation for estimating an original color not show-through from the input image And color adjusting means for performing color adjustment between the color estimated by the color estimating means and the color before estimation according to the edge binary image acquired by the edge binary image acquiring means An image processing apparatus comprising:   The image processing apparatus according to claim 1, wherein the binarization unit binarizes into a background color and other colors by floating binarization processing.   The image processing apparatus according to claim 1, wherein the binarization unit binarizes by a moving average method.   Different from binarization means for binarizing an input image by a moving average method, and after labeling an isolated region in an image obtained by performing shrinkage processing on an image binarized by the binarization means The binarization includes the pixel of interest and the binarization using a binarized image that is subjected to expansion processing so as not to combine the isolated regions to which labels are attached, and a labeled binarized image after processing by the contraction / expansion unit Set by a region setting unit that sets a reference region so that many pixels of a label corresponding to the target pixel are included when the target pixel is labeled in the image, and the region setting unit in the input image An image processing apparatus comprising color estimation means for estimating an original color that is not show-through based on a value of a pixel in the reference area.   Different from binarization means for binarizing an input image by a moving average method, and after labeling an isolated region in an image obtained by performing shrinkage processing on an image binarized by the binarization means The binarization includes the pixel of interest and the binarization using a binarized image that is subjected to expansion processing so as not to combine the isolated regions to which labels are attached, and a labeled binarized image after processing by the contraction / expansion unit When a label is given to the target pixel in the image, a region setting unit that sets a reference region composed of pixels with a label corresponding to the target pixel, and a reference set by the region setting unit in the input image An image processing apparatus comprising color estimation means for estimating an original color that is not show-through based on pixel values in a region.   The image processing apparatus according to claim 8, wherein the color estimation unit performs a color estimation process according to the size of the reference area set by the area setting unit.   From the input image, a binarization unit that binarizes an input image by a moving average method, a contraction / expansion unit that performs an expansion process on the image binarized by the binarization unit, Color estimation means for estimating an original color that is not show-through, and color adjustment between the color estimated by the color estimation means according to the binary image after processing by the contraction and expansion means and the color before estimation An image processing apparatus comprising color adjusting means for performing the above.   An image processing program for causing a computer to execute the function of the image processing apparatus according to any one of claims 1 to 11.

Images