JP5023205B2 - Dropout color processing apparatus and dropout color processing method using the same - Google Patents

Description

  The present invention relates to an extraction image processing apparatus and a processing method for specifying and extracting a color range from multicolor image data and changing the corresponding color to a preset color, and in particular, a color range to be extracted. It is related with the technique of designating with good operability.

  In recent years, the enactment of the e-document law (enforced on April 1, 2005) that legally allows electronic storage (image files) of financial documents has become a trigger, such as OCR (optical character recognition), scanners, etc. Documents are being digitized at offices, government offices, and local governments. In particular, OCR greatly contributes to business efficiency because it can digitize a large amount of handwritten information in a short time. The OCR form input to the OCR has a recognition rate improved by printing a frame, a ruled line, or the like, which is a reference when a human enters a character, in a dropout color and reading only the character written by the human. Conventional character reading was optically extracted by a filter attached to an OCR reading sensor, so it was a function unique to OCR, but now it is digitally extracted by image processing of multicolored images with a computer. Is possible. However, since various OCR forms are independently created, printed, and used for each company or business unit, it is necessary to specify and register the color of the frame to be dropped out according to the environment when introducing OCR.

  However, at the time of organizational reorganization, such as corporate integration or merger of local governments, it is necessary to improve the efficiency of setting changes in order to respond quickly to a wide range of forms from other companies and other local governments before the merger after the introduction of equipment. In addition, multiple dropout colors (multidropout colors) have been added to express a wealth of information in a single form, and it is essential to support the diversity of character colors, including carbon copies of colors. It is expected that improvement of the function for setting the boundary (threshold value) between the character color and the extracted color will be an issue.

  As a method for solving this problem, a method for changing the binarization threshold according to the saturation as in Patent Document 1 and the maximum and minimum values of the color distribution are obtained from a histogram of a specific image portion as in Patent Document 2. A method for obtaining a dropout color range, a method for extracting ruled lines and increasing the separation accuracy from the histogram as in Patent Document 3, and a method for detecting a specific color area from a color distribution in a color space and removing a color background as in Patent Document 4 A method for improving the separation accuracy by using a color distribution to be extracted and a color distribution not to be extracted as in Patent Document 5 has been proposed.

JP 2001-118031 A JP 2003-216894 A JP 2005-258683 A JP-A-9-51443 Japanese Patent Laid-Open No. 7-98762

  These methods or a combination thereof cannot solve the problem that the user's operation of specifying the color to be extracted is complicated and the adjustment of the threshold takes a large amount of man-hours.

  For example, in the method of changing the binarization threshold according to the saturation as in Patent Document 1, it is necessary to set a saturation threshold that distinguishes between chromatic and achromatic colors, and a ruled line with low saturation is dropped out. Needs to be set to a low threshold value, and this causes a basic problem that characters cannot be read in a form in which handwritten characters with high saturation such as a ballpoint pen are written. Further, in the method of obtaining the maximum / minimum value of the color distribution and using the color range as in Patent Document 2 and Patent Document 3, the color extraction / dropout range is rectangular, so that character blurring is prevented. When the extraction area is made smaller, ruled lines appear, and when the extracted area is expanded and the ruled lines are suppressed, characters become blurred. Adjustment of the threshold goes up to the reef.

In addition, in the method of detecting a specific color region from the color distribution in the color space and removing the color background as in Patent Document 4, examples in which various specific color extraction methods are applied to the color plane are shown. Since the extraction is unique to the color plane, it cannot cope with the separation of the portion where the ruled line and the character intersect.
In the method of improving the separation accuracy by using the color distribution to be extracted and the color distribution not to be extracted as in Patent Document 5, the operation of designating the color to be extracted on the screen is performed for each pixel. Or a color close to the center and an allowable width must be input, which is cumbersome, and it is almost impossible to specify characters on a halftone dot background.

  The present invention provides an input unit for inputting a form as multicolor image data in order to solve the problem that a user's operation of specifying a color to be extracted in a dropout color processing apparatus is complicated and takes a lot of man-hours to adjust a threshold value. A color space conversion unit that converts the image data into a predetermined color space, a histogram creation unit that creates a histogram of data in the color space, and a contour creation unit that creates a contour of a data distribution based on the histogram A threshold generation unit that creates an extraction region on the color space based on the created contour, and the image data depending on whether the color space conversion result of each pixel of the image data belongs to the extraction region on the color space. A dropout color processing device having a designated color conversion unit that converts each pixel of the image into a designated color and an output unit that outputs the processed image The threshold generation unit is provided with a plurality of threshold generation units that generate the extraction region in the color space by different methods, and when the contour generation unit generates the contour, whether the coordinates in the color space are within a preset region. The plurality of threshold generation units are switched depending on whether or not.

  In addition, at least one of the plurality of threshold generation units is a threshold generation unit that selects the minimum boundary of the histogram distribution and creates an extraction region, and at least one selects the valley portion of the histogram distribution and selects the extraction region. It is a threshold generation unit to be created.

  Further, the boundary of the region set in advance on the color space is parallel or perpendicular to the color space.

  Further, the boundary of the region set in advance on the color space is provided so as to be inclined with respect to the color space.

An input unit that inputs a form as multicolor image data; a color space conversion unit that converts the image data into a predetermined color space; a histogram creation unit that creates a histogram of data in the color space; A contour creation unit that creates a contour of a data distribution based on a histogram, a threshold generation unit that creates an extraction region in a color space based on the created contour, and the color space conversion result of each pixel of the image data is the color A dropout color having a designated color conversion unit that converts each pixel of the image data into a designated color depending on whether it belongs to an extraction area in space, and an output unit that displays or outputs the processed image In the processing device,
A first range designating unit for designating an image area “including the extracted color” for the image data; a second range designating unit for designating an image area “not containing the extracted color” for the image data; A histogram creation unit that extracts a part or all of the color space distribution in an image area that includes an extracted color and an image area that does not include an extracted color, and a threshold generation that generates a threshold based on the histogram extraction result. It comprises a part.

  In addition, as a means for generating a threshold of “extracted color”, a first smoothing means for generating an “extracted color” envelope surface as a distribution in a color space obtained by color extracting an “extracted color” image area; A second smoothing means for generating an envelope surface that does not include an extracted color as a distribution in a color space obtained by extracting a color of an image area that does not include an extracted color, and at least inside the envelope surface that includes an extracted color And a threshold generation unit having a determination criterion of “extracted color” if it is outside the envelope surface “not including extracted color”.

  In addition, the threshold value generation unit is configured to generate an “extracted color” based on the histogram extraction result in a competitive area that is inside the envelope surface that includes “extracted color” and that is inside the envelope surface that does not include “extracted color”. When creating an `` contained '' contour, select the minimum boundary of the histogram distribution outside the `` does not include extracted color '' area to generate a threshold, and select the valley of the histogram within the `` does not include extracted color '' area It has the threshold value production | generation part which produces | generates a threshold value, It is characterized by the above-mentioned.

  According to the present invention, in a dropout color processing apparatus that determines a predetermined color range and changes to a set color for multicolor image data, a plurality of “extracted color” color ranges and “extracted color” colors Since the range can be set easily and the extraction boundary can be determined automatically, it is easy to add new extraction forms, greatly reduce maintenance costs, and stably extract and remove colors. be able to.

1 is an overall block diagram of a first embodiment of the present invention. It is a schematic flowchart of the 1st Embodiment of this invention. It is a schematic diagram which shows the example of histogram preparation of the 1st Embodiment of this invention. It is a schematic diagram which shows the histogram storage data of the 1st Embodiment of this invention. It is a schematic diagram which shows the threshold value generation of the 1st Embodiment of this invention. It is detailed explanatory drawing which shows the threshold value production | generation system of the 1st Embodiment of this invention. It is detailed explanatory drawing which shows the modification of the threshold value generation system of the 1st Embodiment of this invention. It is a flowchart which shows the threshold value generation of the 1st Embodiment of this invention. It is a schematic diagram which shows the 1st threshold value production | generation system of the 1st Embodiment of this invention. It is a flowchart of the 1st threshold value generation system of the 1st Embodiment of this invention. It is a schematic diagram which shows the 2nd threshold value production | generation system of the 1st Embodiment of this invention. It is a flowchart of the 2nd threshold value generation system of the 1st Embodiment of this invention. It is a schematic flowchart of the 2nd Embodiment of this invention. It is a schematic diagram which shows the threshold value generation of the 2nd Embodiment of this invention. It is explanatory drawing which shows the threshold value generation of the 2nd Embodiment of this invention. It is explanatory drawing which shows the threshold value generation of the 2nd Embodiment of this invention. It is a schematic diagram which shows the operation screen in embodiment of this invention. It is a schematic diagram which shows the other operation screen in embodiment of this invention. It is a block diagram which shows the hardware constitutions of this invention.

Hereinafter, the configuration of the embodiment of the present invention, an image display example, and a processing flowchart will be described.
[First Embodiment]
FIG. 1 is an overall block diagram of the present invention. An input unit 100 for inputting an RGB image performs HSV color space conversion and histogram processing of image data by a gray conversion unit 211, a floating threshold binarization unit 212, and a binarization determination unit 214 in the color space conversion / histogram creation unit 200. Do. Thereafter, the histogram processing unit 213 collects a histogram of only pixels that become black pixels by binarization processing and stores the histogram in the histogram buffer 222. The primary buffer data for binarization is stored in the floating threshold binarized image buffer 220, the image converted from RGB to HSV is stored in the HSV image buffer 230, and the binarization result is stored in the binarization result buffer 240. Store each one.

  Next, the contour creation / threshold value generation unit 300 generates threshold values for the ruled line portion and the character portion by the first threshold value generation unit 310 and the second threshold value generation unit 311 provided therein, and sets an extraction region on the color space. Based on this extraction region, the specified color conversion unit 400 determines whether the color of the target pixel is within the extraction region on the color space, and the threshold determination unit 410 determines the output image from the output unit 110 to an output device such as a display or printer. Output.

  FIG. 2 shows a schematic flowchart of the first embodiment of the present invention. The input image 100a shows a form image in which ruled lines 104 are printed and characters 102 are entered. The color space conversion / histogram creation unit 200 electronically reads a form and converts it into multicolor image data such as RGB, converts it into, for example, an HSV color space close to the characteristics of the human eye, and collects the histogram 201. . Further, the contour creation / threshold generation unit 300 creates a frequency distribution contour from the histogram 201, checks whether there is interference between the created contour and a preset region or the created contour and another contour, and SV A threshold value 301 is generated on the plane. The designated color conversion unit 400 converts the image of the input image data based on the threshold value 301, replaces the ruled line 104 with white and removes it, and outputs an output image 110 a that reproduces only the character 102.

  Next, details of each of the above processes will be described. FIG. 3 is a schematic diagram of the HSV color space. FIG. 3A shows the hue H in the rotation direction of the HSV color space 2000, the saturation S in the system direction, and the luminance V in the height direction. Since the HSV color space is separated into hue, saturation, and luminance as compared to multicolor data such as RGB, processing closer to human recognition is possible. In this embodiment, the hue is classified into 30 types, the saturation and the luminance are classified into 32 types, and the frequency of the image data converted into the HSV color space is obtained. FIG. 3B shows a schematic diagram in which the frequency distribution is obtained from the histogram. The frequency of the character portion such as black ink has a distribution 2002 with low saturation, and the colored ruled lines have distributions 2003 and 2004 with high saturation.

  FIG. 4 is a schematic diagram showing a histogram collection data structure of image data on the HSV color space. Each collected data 2100 to 2104 obtains the frequency of saturation (S) and luminance (V) for each hue (H). Thus, the histogram is stored as three-dimensional data.

  5A and 5B are schematic diagrams of the threshold generation method. FIG. 5A shows a histogram of the HSV color space 2000, and FIG. 5B shows an SV plane in which one hue is cut out. In the figure, the horizontal axis indicates saturation (S), the vertical axis indicates luminance (V), 2005 indicates the distribution of character portions, and 2006 indicates the distribution of ruled line portions. FIG. 5C is an enlarged view of the distribution 2005 of the character part and the distribution 2006 of the ruled line part because there are characters on the ruled line and the distributions are often overlapped. A section 2007 in the overlapped distribution is as follows. For example, as shown in FIG. 5D, valleys are often formed in the frequency distribution. A valley line 2008 is shown in FIG. In the present invention, this frequency valley is used to separate the overlapping distribution of characters and ruled lines.

  FIG. 6 is an explanatory diagram showing an example of the threshold generation process, showing the distribution of characters and ruled lines on the SV plane in a specific hue H. In FIG. 6A, a provisional separation threshold SMin for characters 2010 and ruled lines 2011 is provided on the horizontal axis S, and a threshold approximate straight line 2012 is obtained for a histogram having a saturation greater than SMin. Next, the intersection of this approximate straight line and the straight line S = SMin is obtained. For the portion of saturation smaller than SMin, the valley of the frequency of the character portion and the ruled line portion described with reference to FIG. Regarding the specific calculation of the threshold, the first threshold generation method is used for a histogram with a saturation larger than SMin, and the second threshold generation method is used for a histogram with a saturation smaller than SMin. It will be described later.

  As a threshold for color separation, a threshold 2013 is generated by applying a predetermined correction coefficient to a broken line composed of a threshold approximate line and a valley line obtained by the above calculation. VMin on the vertical axis V indicates the upper limit of noise, and threshold calculation is only for VMin and higher.

  When the character frequency distribution is smaller than the temporary separation threshold SMin, the characters and ruled lines can be correctly separated as described above. On the other hand, when the character distribution 2016 is larger than the temporary separation threshold SMin as shown in FIG. 6B, an error occurs between the threshold approximate line and the actual distribution, and the threshold is set low. A distribution 2017 occurs above the threshold value, causing a problem that the read characters are faint.

FIG. 7A shows another temporary separation threshold setting method corresponding to the above. The temporary separation threshold 2020 is set as a straight line that is not parallel to the S-axis and is inclined approximately 45 degrees. Thereby, the deviation between the temporary separation threshold and the actual distribution can be reduced, and the robustness of threshold generation can be improved. Further, in a region of saturation greater than the temporary separation threshold, the histogram is not a linear approximation but the minimum frequency boundary 2021 is obtained. As a result, the correction coefficient for converting the threshold approximation line into the threshold value can be eliminated, and the number of parameters can be reduced. FIG. 7B shows a threshold value determination method for an area smaller than the temporary separation threshold value 2020, which uses a method for obtaining a trough portion of the frequency of characters and ruled lines.

  FIG. 8 shows an overall flowchart of the threshold value generation process of the present invention. First, variables and tables are initialized (S11), and coordinates are updated (S12). For coordinate movement, for each hue H, the color frequency is scanned in descending order and the luminance V is scanned in ascending order from the noise upper limit VMin. As described in FIGS. 6 and 7, when the target coordinates are outside the competitive area, that is, when the target saturation is higher than the temporary separation threshold (S13), the minimum value extraction (S14) which is the first threshold generation method is performed. If the saturation of the target is lower than the temporary separation threshold (S13), the boundary is obtained by frequency valley detection (S21) which is the second threshold generation method. After the boundary calculation, the previous value is updated (S15) to store the previous V position. If the scan is completed (S16), the process proceeds to the next step. If the scan remains, the process returns to the coordinate update (S12). If the saturation S is processed in descending order and the processing is completed (S16), the coordinates of the saturation S are updated in ascending order (S17), and the smoothing process (S18) is performed to convert the sudden change into a smooth motion. . If the scan is completed (S19), the threshold generation is completed (S20).

Next, two threshold value generation methods used in the embodiment of the present invention will be described in detail.
FIG. 9 is a schematic diagram for explaining the details of the minimum value detection method which is the first threshold value generation method. FIG. 9A shows the frequency on the SV plane in the hue H. FIG. H1 to H7 are obtained by connecting the minimum value and the maximum value of the luminance V with a frequency exceeding a predetermined threshold by a line. Since the frequency distribution depends on the image, it does not always exist at all saturations S, and here H2 and H5 are missing. In FIG. 9B, the minimum value of the brightness V corresponding to the saturation S is obtained in descending order of the saturation S from the maximum saturation. This flowchart is shown in FIG. When the current minimum value (current V) is smaller than the previously obtained minimum value (previous V) (S33), that value is adopted as it is (S34). However, when the current V is larger than the previous V (S33) or when there is no frequency, the previous V is increased by 1 (S35), and a steep change is smoothed (FIG. 9C). When this process ends, the saturation S is increased in ascending order from 0, and when the current V is larger than the previous V, the previous V is increased by 1 to complete smoothing to both (FIG. 9 (d) )). Smoothing is performed in this way, and an extraction region desired by the user is cut out from the frequency obtained from the histogram.

FIG. 11 is a schematic diagram for explaining a frequency valley detection method which is the second threshold generation method. The horizontal direction represents saturation S, and the vertical direction represents luminance V. P0, P1, and P2 indicate the circumstances for determining the saturation S in descending order. When obtaining the next threshold boundary from P2, it is determined which of P4, P5 and P6 is to be proceeded to. The flowchart is shown in FIG. The frequencies of the reference pixel A and the reference pixel B are compared (S43). If the frequency of A is smaller than the frequency of B, the process proceeds to P4 (S44), and if the frequency of A is equal to the frequency of B, the process proceeds to P5 ( If the frequency of A is greater than the frequency of B (S45), the process proceeds to P6 (S46). In this way, it is possible to follow the valley of the frequency and determine the threshold value.
[Second Embodiment]
Next, a second embodiment of the present invention will be described. In the first embodiment, a temporary separation threshold is set in the color space in order to extract characters close to an achromatic color such as black described on a colored ruled line.
In the second embodiment, in order to cut out and extract characters written in chromatic colors such as color ballpoint pens, magic, and color carbon, the first non-existing color area is used as a temporary separation threshold. The threshold value generation process can be executed as in the embodiment.

A schematic flowchart of the present embodiment will be described with reference to FIG. When extracting color characters, it is necessary to define which color to extract using a reference image. In the reference image 101a, characters of the object 102 to be extracted and mixed ruled lines are mixed. This is instructed by the first range designating unit that designates the range 105 that includes “extracted color” and the second range designating unit that designates the range 103 that does not include the extracted color. Is determined. For this reason, it is not necessary to designate one dot in a fine character to be extracted as in the conventional example, and it is not necessary to input a margin threshold value for the extracted center color.

  Histograms are generated by the histogram creation units 251 and 252 for the “excluded color” range 103 and the “extracted color” range 105, respectively, and the “excluded color” histogram 261 and the “excluded color” histogram 262 is obtained. A character extraction threshold value 301 is created from the plurality of histograms by the contour formation / threshold value generation unit 300a. The contour formation / threshold generation unit 300a determines that the image data is “extracted color” if the image data is at least inside the envelope surface “including the extracted color” and outside the envelope surface “not including the extracted color”. Has a standard. Based on this threshold value, the dropout process is executed by converting the extracted color into a designated color based on whether the HSV value of each pixel of the input image 100 is within the extraction boundary on the color space.

  FIG. 14 is a schematic diagram of threshold generation. FIGS. 14C and 14D show cross sections cut out with a specific hue H in the histogram including the extracted color in FIG. 14A and the histogram in FIG. 14B not including the extracted color. In the first embodiment, a temporary separation threshold is set in advance in the color space. In the second embodiment, the first embodiment uses an area that does not include an extracted color as the temporary separation threshold 2021. A threshold generation process similar to that in the embodiment can be executed. That is, in FIG. 14E, if it is outside the temporary separation threshold 2021, the threshold is generated by the minimum value detection that is the first threshold generation method described above, and if it is within the temporary separation threshold 2021, the second is detected. The threshold value is generated by detecting the valley of the frequency that is the threshold generation method. FIG. 14F is a schematic diagram illustrating a frequency trough portion in the cross-section 2007 in FIG.

  FIG. 15 is a schematic diagram specifically illustrating the threshold generation method according to the present embodiment. 2011 is a ruled line distribution, and 2016 is a character distribution. Processing is performed by the first smoothing means for generating an envelope surface including the extracted color. That is, FIG. 15A shows a schematic diagram of detection of the minimum value 2041 outside the area 2045 not including the extracted color. The saturation S is detected in descending order, and the minimum value of the extraction area is detected while being smoothed. Next, in FIG. 15B, the maximum value 2042 is detected outside the area 2045 that does not include the extracted color.

Further, in FIG. 16, processing is performed by the second smoothing unit that generates an envelope surface that does not include the extracted color, that is, the frequency valley line 2043 is detected in the area 2045 that does not include the extracted color.
When these detection results are connected, an extraction region 2046 desired by the user can be automatically obtained as shown in FIG. 16B, and stable color extraction can be performed with simple designation without combining complicated threshold values. Is possible.

  FIG. 17 shows a screen display example of the present invention. The screen 3000 includes a screen preview 3001, a histogram display 3002, and an edit button 3010. When the color designation button 3003 is ON, display is performed along a conventionally used interface. The read reference image is displayed on the preview screen 3001, and it is determined whether the extracted pixel button 3011 or the unextracted radio button 3012 is ON as to whether the pixel is to be extracted or not. To give instructions.

  When a pixel that is extracted or not extracted is clicked on the preview screen 3001, the center color space is displayed in the color space on the histogram display 3002, and the range can be defined by adjusting the bounding box 3022 with the mouse. When the extraction / non-extraction area is determined, editing processing can be performed with buttons such as color range registration 3008, addition 3006, and deletion 3009.

  FIG. 18 shows another image display example of the present invention. When the inclusion designation button 3003a is ON, the range 105 including the extracted color and the range 103 not including the extracted color can be specified on the image preview 3001. This is effective when ruled lines and characters are mixed in a small area or when characters on a halftone dot background are designated. Whether or not to include the extracted color is instructed by ON / OFF of the radio button 3011 to be extracted and the radio button 3012 not to be extracted in the same manner as described above.

FIG. 19 shows a hardware configuration example for realizing the present invention. A color image is input from the scanner 14, and after the image is extracted by the personal computer 15, the image is transferred to the recognition device 17.
Reference numeral 151 denotes a CPU, 152 denotes an image control unit, 153 denotes a USB connection unit, 154 denotes a memory, 155 denotes a LAN I / F, 156 denotes an ATA I / F, and 157 denotes a hard disk. Setting of the threshold value and the color range is instructed by using a keyboard 18 and a mouse 19 connected to the personal computer 15.

  The present invention can be applied not only to a form reading apparatus but also easily to computer image processing software, OCR software, and the like.

  Other embodiments possible in the present invention are described below.

  (Supplementary Note 1) An input unit that inputs a form as multicolor image data, a color space conversion unit that converts the image data into a predetermined color space, and a histogram creation unit that creates a histogram of data in the color space; A contour creation unit that creates a contour of a data distribution based on the histogram, a threshold generation unit that creates an extraction region in a color space based on the created contour, and the color space conversion result of each pixel of the image data A designated color conversion unit that converts each pixel of the image data into a designated color depending on whether or not it belongs to an extraction region on the color space, and an output unit that displays or outputs the processed image, A first range designating unit for designating an image area “including an extracted color” for the image data; and a second range designating unit for designating an image area “not including an extracted color” for the image data; A histogram creation unit that extracts part or all of the color space distribution in the image area that includes the extracted color and the image area that does not include the extracted color, and a threshold value that generates a threshold based on the histogram extraction result Dropout color processing apparatus characterized by comprising a generation unit In the dropout color processing apparatus, as means for generating a criterion for “extracted color”, the “extracted color is included” inside the envelope surface, A dropout color processing apparatus comprising means for changing a degree of extraction in a competing region inside an envelope surface that does not include color according to a distance between boundary colors of the envelope surface.

  (Supplementary Note 2) In the dropout color processing apparatus according to Supplementary Note 1, means for designating an image area “including the extracted color” for the image data or designating an image area “not including the extracted color” for the image data Means for presenting input image data or binarized data to the operator as a means, and whether the operator designates an image area “includes extracted color” or “does not include extracted color” Means for switching between the above designations, means for changing the color of the rectangle displayed after dragging and designation according to the designation of “including extracted color” or “not including extracted color”, and after A dropout color processing apparatus comprising means for storing a start / end position of an image area.

  (Supplementary note 3) The dropout color processing apparatus according to supplementary note 1, further comprising means for displaying a criterion for determining “extracted color” on a two-dimensional or three-dimensional frequency distribution image.

  (Supplementary note 4) In the dropout color processing apparatus according to supplementary note 3, means for displaying an output image generated based on a plurality of “extracted color” determination criteria in a different color, and a target image relative to the output image Dropout color processing characterized by having means for designating a color corresponding to an area and target output, means for instructing the correct answer of the extracted location, means for displaying correct data, and means for registering correct data apparatus.

  (Additional remark 5) In the dropout color processing apparatus according to additional remark 4, an edit means for adding or deleting a new “extracted color” to the already set “extracted color” criterion is performed. As a result, a dropout color processing apparatus having means for displaying a difference from an output based on a criterion before editing.

  (Supplementary note 6) The dropout color processing apparatus according to supplementary note 5 includes means for displaying a difference from the correct answer data in addition to the output image before editing and the output image after editing of the criterion of “extracted color”. Dropout color processing device.

  (Supplementary note 7) An input unit that inputs a form as multicolor image data, a color space conversion unit that converts the image data into a predetermined color space, and a histogram creation unit that creates a histogram of data in the color space; A contour creation unit that creates a contour of a data distribution based on the histogram, a threshold generation unit that creates an extraction region in a color space based on the created contour, and the color space conversion result of each pixel of the image data A designated color conversion unit that converts each pixel of the image data into a designated color depending on whether or not it belongs to an extraction region on the color space, and an output unit that displays or outputs the processed image, A first range designating unit for designating an image area “including an extracted color” for the image data; and a second range designating unit for designating an image area “not including an extracted color” for the image data; A histogram creation unit that extracts part or all of the color space distribution in the image area that includes the extracted color and the image area that does not include the extracted color, and a threshold value that generates a threshold based on the histogram extraction result In a dropout color processing apparatus including a generation unit, as means for generating a threshold of “extracted color”, an envelope surface including “extracted color” as a distribution in a color space obtained by performing color extraction on an image region including “extracted color” And a second smoothing means for generating an envelope surface that does not include an extracted color as a distribution on a color space obtained by color extracting an image area that does not include an extracted color, and at least Means for generating an envelope surface, comprising a threshold generation unit having a criterion of being “extracted color” if it is inside the envelope surface “including the extracted color” and outside the envelope surface “not including the extracted color” Dropout color processing characterized by having a means for detecting the minimum and maximum luminance values for each hue and for each saturation, performing smoothing processing in the saturation direction, and further performing smoothing processing in the hue direction apparatus.

  (Supplementary Note 8) An input unit that inputs a form as multicolor image data, a color space conversion unit that converts the image data into a predetermined color space, and a histogram creation unit that creates a histogram of data in the color space; A contour creation unit that creates a contour of a data distribution based on the histogram, a threshold generation unit that creates an extraction region in a color space based on the created contour, and the color space conversion result of each pixel of the image data A designated color conversion unit that converts each pixel of the image data into a designated color depending on whether or not it belongs to an extraction region on the color space, and an output unit that displays or outputs the processed image, A first range designating unit for designating an image area “including an extracted color” for the image data; and a second range designating unit for designating an image area “not including an extracted color” for the image data; A histogram creation unit that extracts part or all of the color space distribution in the image area that includes the extracted color and the image area that does not include the extracted color, and a threshold value that generates a threshold based on the histogram extraction result In a dropout color processing apparatus including a generation unit, as means for generating a threshold of “extracted color”, an envelope surface including “extracted color” as a distribution in a color space obtained by performing color extraction on an image region including “extracted color” And a second smoothing means for generating an envelope surface that does not include an extracted color as a distribution on a color space obtained by color extracting an image area that does not include an extracted color, and at least Means for generating an envelope surface, comprising a threshold generation unit having a criterion of being “extracted color” if it is inside the envelope surface “including the extracted color” and outside the envelope surface “not including the extracted color” A dropout color processing apparatus comprising means for detecting a minimum value and a maximum value of saturation for each hue and for each luminance, performing a smoothing process in the luminance direction, and further performing a smoothing process in the hue direction .

100: input unit 110: output unit 200: color space conversion / histogram creation unit 300: contour creation / threshold generation unit 300a: contour creation / threshold generation unit 310: first threshold generation unit 311: second threshold generation unit 400: designation Color converter

Claims (6)

An input unit that inputs a form as multicolor image data, a color space conversion unit that converts the image data into a predetermined color space, a histogram creation unit that creates a histogram of data in the color space, and the histogram An outline creation unit for creating a contour of the data distribution based on the threshold, a threshold generation unit for creating a threshold value based on the created outline and creating an extraction region on the color space, and the color space conversion result of each pixel of the image data Dropout color processing having a designated color conversion unit that converts each pixel of the image data into a designated color depending on whether or not it belongs to an extraction region on the color space, and an output unit that outputs the processed image In the device
A first range designating unit for designating an image area “including an extracted color” for the image data; a second range designating unit for designating an image area “not containing an extracted color” for the image data; A histogram creating unit that extracts a part or all of the color space distribution in the image area including the color and the image area not including the extracted color, and the image in the color space distribution based on the histogram extraction result. A threshold generation unit that generates a threshold based on a determination criterion for determining “extracted color” if data is inside a histogram envelope that includes “extracted color” and outside a histogram envelope that does not include “extracted color” A dropout color processing apparatus comprising:   2. The dropout color processing apparatus according to claim 1, wherein, as means for generating a threshold value of “extracted color”, an envelope surface including “extracted color” as a distribution in a color space obtained by color-extracting an image region including “extracted color” First smoothing means for generating an image plane, a second smoothing means for generating an envelope surface that does not include an extracted color as a distribution in a color space obtained by extracting the color of an image area that does not include an extracted color, and at least an “extracted color” A dropout color process comprising: a threshold generation unit having a determination criterion of being “extracted color” if it is inside the envelope surface “including” and outside the “excluded color” envelope surface apparatus.   The dropout color processing apparatus according to claim 1, wherein the threshold value generating means is inside the envelope surface “including the extracted color” and inside the envelope surface “not including the extracted color”, When creating an outline that includes “extracted color” based on the above histogram extraction result, a threshold is generated by selecting the minimum boundary of the histogram distribution outside the “excluded color” area, and an “excluded color” area A dropout color processing apparatus comprising a threshold value generation unit that generates a threshold value by selecting a valley portion of the histogram. A form is input as multicolor image data, the image data is converted into a predetermined color space, a histogram of the data in the color space is created, a contour of the data distribution is created based on the histogram, and the creation A threshold value is generated based on the contour, an extraction area on the color space is created, and each pixel of the image data is determined depending on whether the color space conversion result of each pixel of the image data belongs to the extraction area on the color space. In the dropout color processing method that converts each to the specified color and outputs the processed image,
For the image data, an “extracted color” image region and an “extracted color” image region are designated, and one of the “extracted color” image region and the “extracted color” image region is specified. Each or all of the color space distributions are extracted to create respective histograms, and based on the histogram extraction result, the image data in the color space distribution is inside the histogram envelope surface including “extracted colors” and “ A dropout color processing method, wherein a threshold value is generated according to a determination criterion for determining “extracted color” if it is outside a histogram envelope surface that does not include an extracted color .   5. The dropout color processing method according to claim 4, wherein an envelope surface including “extracted color” is used as a distribution in a color space obtained by color extracting an image region including “extracted color” in order to generate a threshold value of “extracted color”. Generate an envelope surface that does not include the extracted color as a distribution in the color space obtained by color-extracting the image area that does not include the extracted color, and that is at least inside the envelope surface that includes the extracted color. A dropout color processing method, wherein if it is outside an envelope surface that does not include an extracted color, it is determined that it is an “extracted color”.   5. The dropout color processing according to claim 4, wherein the threshold value generation unit is arranged in the competitive region that is inside the envelope surface that includes the “extracted color” and that is inside the envelope surface that does not include the “extracted color”. When creating an outline that includes “extracted color” based on the histogram extraction result, a threshold is generated by selecting the minimum boundary of the histogram distribution outside the “excluding extracted color” area, and within the “excluding extracted color” area Then, a dropout color processing method, wherein a threshold value is generated by selecting a valley portion of a histogram.

Images