JP2005217984A - Image processing device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing device eliminating ground portions of image data while giving consideration to an image with low chroma saturation. <P>SOLUTION: From among the image data to be processed, a ground region is extracted to produce a color correction parameter based on color information relating to a color element except for luminance represented by pixels within the ground region extracted. By using the color correction parameter produced, the image data to be processed are corrected in color to eliminate the ground portions. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus that performs processing for removing a background portion of image data.

  When processing image data read by a scanner such as a copying machine, it is preferable to remove the background portion in advance. That is, the image data read by the scanner is often colored on, for example, a paper portion that is originally white, and the background color often appears.

  Conventionally, this background removal processing has been performed as follows. That is, the processing functionally includes a lightness histogram calculation unit 1, a background threshold value determination unit 2, an LUT creation unit 3, and a lightness correction unit 4, as shown in FIG.

  The brightness histogram calculation unit 1 calculates a brightness histogram based on the image data read from the scanner. The background threshold value determination unit 2 searches the calculated histogram from the low (dark) lightness side and acquires a lightness value having a frequency exceeding a predetermined threshold value. This brightness value is assumed to be THD1.

を生成する。なお、これらのパラメータは任意のＴＨＤ１について、ｋ×ＴＨＤ１＋ｃ＜ｄ×ＴＨＤ１となるように設定されているものとする。 The LUT generation unit 3 uses predetermined parameters k, c, and d,

Is generated. Note that these parameters are set so that k × THD1 + c <d × THD1 for any THD1.

  The LUT generation unit 3 outputs 0 for input values from 0 to THD1 ′, x for input values x such that x ≧ THD2, and THD1 ′ ≦ x <THD2 As for LUT, an LUT (FIG. 7) that generates an output value corresponding to a straight line connecting coordinates (THD1 ′, 0) to (THD2, THD2) is generated.

  The brightness correction unit 4 corrects the brightness of the image data read from the scanner based on the LUT generated by the LUT generation unit 3.

  For example, when a newspaper is read and the background portion is gray, the histogram is as shown in FIG. A corresponding LUT is generated to correct the image data. By this correction, the white color of the paper surface is reproduced, and the corrected histogram is as shown in FIG.

A technique related to background removal using a method similar to this method is disclosed in Patent Document 1.
JP-A-4-37258

  However, although the conventional background removal method exhibits sufficient performance for black-and-white documents, it may be difficult to see low-saturated images and the like by applying the LUT to colored documents and the like. . Further, in the case of image data printed on colored paper, the color of the paper may not be reproduced.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide an image processing apparatus capable of removing a background portion of image data while considering an image with low saturation.

  The present invention for solving the problems of the above-described conventional example is an image processing apparatus that performs a process of removing a background portion that is a background of a significant image from image data that is a processing target, and the processing target image Means for extracting a background area from the data; means for generating a color correction parameter based on color relationship information related to a color component other than brightness represented by a pixel in the extracted background area; and the generation And correcting means for correcting the color of the image data to be processed using the color correction parameter.

  Further, here, the color-related information is information on saturation expressed by pixels in the background area, the color correction parameter is a parameter for correcting saturation, and the means for correcting the color includes: The saturation of each pixel included in the image data to be processed may be corrected using the color correction parameter.

  In these cases, the lightness may be corrected using the same lightness LUT as in the past, but the parameter for correcting the lightness based on the information related to lightness represented by the pixels in the extracted background region. The correction unit may correct the brightness component of the image data to be processed using the parameter for correcting the brightness and the color correction parameter.

  Furthermore, the color relationship information may be information relating to the distance between the pixel value in the base region in the color space and the specific pixel value. At this time, the information on the distance may be a distance defined by assigning different weights to each of the brightness component and the other components.

  In addition, the present invention for solving the problems of the conventional example is an image processing method for performing processing for removing a background portion as a background of a significant image from image data to be processed using a computer. The color correction parameter is generated based on the color relation information related to the color component other than the brightness represented by the pixel in the extracted background area and the step of extracting the background area from the image data to be processed And a step of correcting the color of the image data to be processed using the generated color correction parameter.

  Furthermore, an image processing program according to another aspect of the present invention provides a computer to perform processing for removing a background portion that is a background of a significant image from image data to be processed. A procedure for extracting a background area from data, a procedure for generating a color correction parameter based on color relationship information related to a color component other than lightness represented by a pixel in the extracted background area, and the generation The computer is caused to execute a procedure for correcting the color of the image data to be processed using the color correction parameter.

  Embodiments of the present invention will be described with reference to the drawings. As shown in FIG. 1, the image processing apparatus according to the first embodiment of the present invention includes a control unit 11, a storage unit 12, an image input unit 13, and an image output unit 14.

  The control unit 11 operates according to a program stored in the storage unit 12, performs a background removal process on the image data input from the image input unit 13, and outputs it from the image output unit 14. Yes. Specific contents of the operation of the control unit 11 will be described later. The storage unit 12 stores a program executed by the control unit 11, and also operates as a work memory that holds various data generated during the processing of the control unit 11.

  The image input unit 13 is connected to an external device such as a scanner, and outputs image data input from the external device to the control unit 11. The image output unit 14 outputs image data to an external device in accordance with an instruction input from the control unit 11.

  Here, specific contents of the background removal processing by the control unit 11 will be described. As shown in FIG. 2, the background removal process performed by the control unit 11 functionally includes a color space conversion unit 21, a brightness feature value calculation unit 22, a brightness background threshold value determination unit 23, and a background region extraction unit. 24, a lightness correction LUT generation unit 25, a saturation feature value calculation unit 26, a saturation correction LUT generation unit 27, and a correction unit 28.

の数式によって行うことができる。なお、明度成分が独立している色空間としては、他にＬａｂ、ＨＳＶなどがあるが、これらのどれであってもよい。 The color space conversion unit 21 converts the color space of the image data input from the image input unit 13 into a color space suitable for processing. In general, image data input from a scanner or the like is expressed in an RGB color space, whereas in the present embodiment, brightness and other color components are handled separately. It is preferred that the components are represented in independent color spaces. Therefore, the color space conversion unit 21 converts, for example, the color space of the image data into a color space such as YCbCr. This conversion is

It can be performed by the following formula. In addition, examples of the color space in which the brightness component is independent include Lab and HSV, and any of these may be used.

  The brightness feature amount calculation unit 22 generates a histogram for the brightness component in the image data output from the color space conversion unit 21. On the histogram, the lightness background threshold value determination unit 23 searches for the lightness value having a frequency exceeding a predetermined threshold value by searching from the low lightness (dark) side. Hereinafter, this brightness value is referred to as THD1. This process is similar to the process of the conventional background threshold value determination unit 2.

  The background area extraction unit 24 outputs, as image data of the background area, a mask image that specifies a portion of the image data output from the color space conversion unit 21 whose brightness value is lower than THD1 (that is, brighter than THD1). .

を生成する。なお、これらのパラメータは任意のＴＨＤ１について、ｋ×ＴＨＤ１＋ｃ＜ｄ×ＴＨＤ１となるように設定されているものとする。 The lightness correction LUT generation unit 25 uses predetermined parameters k, c, and d,

Is generated. Note that these parameters are set so that k × THD1 + c <d × THD1 for any THD1.

  The lightness correction LUT generator 25 outputs 0 for input values from 0 to THD1 ′, x for input values x such that x ≧ THD2, and THD1 ′ ≦ x. For THD2, an LUT (FIG. 7) is generated that has an output value corresponding to a straight line connecting coordinates (THD1 ′, 0) to (THD2, THD2). This process is the same as the process in the conventional LUT generation unit 3.

  The saturation feature amount calculation unit 26 is a predetermined saturation feature amount from the image data of the part specified by the image data of the background region output from the background region extraction unit 24 in the image data output from the color space conversion unit 21. Is calculated. Specifically, among the pixel values of each pixel included in the specified portion, a value related to saturation is calculated based on a color component other than lightness (here, Cb and Cr color difference components). Calculate and output as a feature amount C. Here, the value related to the saturation can be, for example, the sum of squares of Cb and Cr (or the square root thereof). Further, the saturation feature amount calculation unit 26 outputs the maximum value (hereinafter referred to as Cmax) among the values related to the saturation for each pixel in the background region together with the saturation feature amount. .

  Further, the saturation feature amount may be a statistic based on the amount related to the saturation for a plurality of pixels. The statistic here can be, for example, a value smoothed by N × N pixels, or a minimum value, maximum value, median value, or the like in N × N pixels. In this way, the influence of isolated point noise can be suppressed.

  The saturation correction LUT generation unit 27 generates an LUT (saturation correction LUT) for correcting the saturation using the saturation feature amount. Specifically, the saturation correction LUT generation unit 27 uses the saturation feature amount calculation unit 26 as a saturation feature amount of an amount related to the saturation related to the pixels in the background area (here, saturation itself). When the average value Cave is output and the maximum value Cmax is output together with it, the saturation correction LUT is generated as follows.

として、彩度の入力値ｘに対して、

となるようなＬＵＴを生成する。ただしＬはＬ≧０、Ｐ＞Ｌなる値であって、通常はＬ＝０，Ｐ＝ＣＴＨ２とすればよい。 That is,

For an input value x of saturation,

LUT is generated as follows. However, L is a value such that L ≧ 0 and P> L, and normally L = 0 and P = CTH2.

  An example of this LUT is shown in FIG. In general, when a low-saturation image is included in a part of the paper, and the low-saturation image portion is mistaken as a background by the processing of the background region extraction unit 24, the area of the low-saturation image portion Is estimated to be small from the area of the entire substrate. If so, the average saturation value Cave of the background portion is close to the saturation of the true background (pixels other than the pixels (significant pixels) constituting the image). On the other hand, the maximum saturation value Cmax is close to the saturation of the erroneously recognized low saturation region. Therefore, here, the average saturation value Cave is set as the first threshold value, the average (Cave + Cmax) / 2 is set as the second threshold value, and the saturation is set to L (for example, “ 0 "), and from the first threshold value to the second threshold value (saturation value lower than the saturation value of the low-saturated region that was misidentified and higher than the saturation value of the background to be removed) It is set so that the chroma value is continuously changed from L to P (for example, a value equal to the second threshold value), and then the saturation value is not changed in the portion exceeding the second threshold value. That is, the first threshold value CTH1 and the second threshold value CTH2 are not limited to the example shown by the expression (2).

  Here, it is also preferable that the value L on the LUT can be set by the user. This is because the saturation of the background portion is determined by the value of L. For example, when light yellow paper is used, if this L is set to the light yellow saturation value, the background removal is performed. Paper color can be maintained while processing.

  For the image data output from the color space conversion unit 21, the correction unit 28 uses the LUT (saturation correction LUT, color correction parameter of the present invention) generated by the saturation correction LUT generation unit 27 for the saturation component. The saturation is corrected based on the Here, as a saturation correction method, each pixel is sequentially selected as a correction target, and the saturation (calculated in the same manner as in the saturation feature amount calculation unit 26) Cin of the pixel to be corrected is calculated. Conversion is performed by the saturation correction LUT, and a conversion result Cout is obtained. Then, the change ratio, k = Cout / Cin, is calculated.

と補正する。 The correction unit 28 then corrects the Cb and Cr components of the pixel value to be corrected.

And correct.

とする。なお、ここでは直線的な重みづけとしているが、ｋ＝０でＹ_out＝Ｙ，ｋ＝１でＹ_out＝Ｙ_LUTとなるような重みづけであれば他の重みづけ方法であってもよい。 Further, for the lightness Y, the correction unit 28 temporarily corrects the lightness based on the LUT generated by the lightness correction LUT generation unit 25 (lightness correction LUT, which corresponds to the parameter for correcting the lightness of the present invention). Temporary correction lightness Y_LUT is generated. Then, a weighted average according to the change ratio of the saturation value is calculated between the lightness Y_LUT after the temporary correction and the lightness Y before the correction to calculate the lightness Y_out after the correction. That is,

And Note that although linear weighting is used here, other weighting methods may be used as long as k = 0, Y_out = Y, k = 1, Y_out = Y_LUT.

  In this way, if only the lightness is determined, even in a low saturation portion that is mistaken for the background region, the saturation of the portion is not lowered, and the first and second saturations are not related. Since the lightness and saturation are continuously corrected between the threshold values CTH1 and CTH2, no false color portion due to image processing occurs at the boundary between the low saturation portion and the background.

  Further, in the above description, the color-related correction is performed using the saturation information. Instead, the color-related correction is performed using a distance in a color space from a specific pixel value (for example, white). Since correction may be performed, an image processing apparatus according to the second embodiment of the present invention that performs such correction will be described below.

  The image processing apparatus according to the present embodiment has substantially the same configuration as that of the image processing apparatus according to the first embodiment shown in FIG. 1, but the processing in the control unit 11 is different. That is, the processing of the control unit 11 in the present embodiment functionally includes a color space conversion unit 21, a brightness feature value calculation unit 22, a brightness background threshold value determination unit 23, and a background area, as shown in FIG. The extraction unit 24, the brightness correction LUT generation unit 25, the distance plane calculation unit 31, the distance feature amount calculation unit 32, the correction parameter generation unit 33, and the correction unit 28 are configured. In addition, about the part which performs the operation | movement similar to the process of the control part 11 in 1st Embodiment, the same code | symbol is attached | subjected and the detailed description is abbreviate | omitted.

  The distance plane calculation unit 31 sequentially selects each pixel value of the image data output from the color space conversion unit 21 as a target pixel value, and calculates a distance between the target pixel value and a predetermined specific pixel value. And output. Here, this distance is calculated as a distance defined by assigning different weights to the brightness component of the target pixel value and the other components.

  Specifically, for the pixel value (Y, Cb, Cr), the distance w when the specific pixel value is white (0, 0, 0) is determined as follows.

  Here, ky is a weight for a lightness component, kc is a weight for a component other than lightness, and when this distance is used for correction relating to color as in this embodiment, as a component related to color, other than lightness The weight is set so as to increase the weight of the component. For example, ky = 1 and kc = 2.

  The distance feature quantity calculation unit 32 stores the distance from the specific pixel value output for each target pixel from the distance plane calculation unit 31 in the storage unit 12 in association with the position information of the target pixel. Specifically, the distance feature amount calculation unit 32 secures an area for storing image data having the same size as the original image data in the storage unit 12, and sets the target pixel at a position corresponding to the target pixel in the area. What is necessary is just to store the value of the distance which the distance plane calculation part 31 outputs.

  The distance feature amount calculation unit 32 receives the input of the mask image data specifying the background region from the background region extraction unit 24, and based on the distance information regarding each pixel specified as the background region, the statistical value (For example, average value, minimum value, maximum value, median value, standard deviation, histogram peak value, etc.) are calculated and output as distance feature amounts.

  The correction parameter generation unit 33 first generates an LUT. Specifically, the average value Wave and the standard deviation Wstd of the distance feature value calculated by the distance feature value calculation unit 32 (that is, the average and standard deviation of the distance from the specific pixel value for the pixel value included in the background region). Is used to generate the following LUT. That is, the converted distance X is determined as follows with respect to the distance x from the predetermined pixel value.

なお、第１の距離しきい値ＸＴＨ１及び、第２の距離しきい値ＸＴＨ２は、それぞれ、

とすることができ、各しきい値での変換後の距離値ＸＴＨ１_New及びＸＴＨ２_Newについては、

とすることができる。ここでｒ０は「０」としてもよい。

The first distance threshold value XTH1 and the second distance threshold value XTH2 are respectively

For the distance values XTH1_New and XTH2_New after conversion at each threshold,

It can be. Here, r0 may be “0”.

  That is, since the color is light when the distance is less than the standard deviation from the average distance from the specific pixel value in the portion extracted as the background area, the distance is set to 0 and the same value as the specific pixel value. To. Further, where the distance exceeding the standard deviation from the average distance from the specific pixel value is exceeded, the original distance value is used as it is so as not to change the color.

と定めて、補正部２８に出力する。 Then, the correction parameter generation unit 33 uses the LUT (distance LUT) generated in this way to generate a parameter k (n, m) for correcting components other than brightness for each pixel. Here, (n, m) is coordinate information representing the position of the pixel to which the parameter is to be applied. Specifically, when the distance between the pixel value of the pixel at the position (n, m) and the specific pixel value is x (n, m), the converted value of the distance, x_New (n, m) is used. , The correction parameter

And output to the correction unit 28.

  The correction unit 28 receives the input of the correction parameter, and the component other than the lightness among the pixel values (Y_out, Cb_out, Cr_out) after correcting the pixel values (Y, Cb, Cr) at the position (n, m). Is calculated as follows.

とする。 Further, with respect to the lightness, the lightness is temporarily corrected based on the LUT generated by the lightness correction LUT generation unit 25 (lightness correction LUT, which corresponds to a parameter for correcting the lightness of the present invention), and a temporary correction lightness Y_LUT is generated. . Then, a weighted average corresponding to the change ratio of components other than lightness (that is, the correction parameter here) is calculated between the lightness Y_LUT after the temporary correction and the lightness Y before the correction, and the lightness Y_out after the correction is calculated. Calculate. That is,

And

  Specifically, by setting the specific pixel value here to be white, if the determination is based only on the lightness by these processes, the saturation of the portion is reduced even in a low-saturation portion that is mistaken for the background region. In addition, since the brightness and saturation are continuously corrected between the first and second distance threshold values XTH1 and XTH2 regarding the distance in the color space from white, the low saturation portion and the background No false color portion associated with image processing occurs at the boundary between the two. Further, the brightness state can be reflected also in the correction relating to the saturation.

  In the above description, the example in which the image data that has undergone the background removal processing is output to an external device (for example, a display, a printer, a disk device, etc.) has been described. The image data may be processed.

  An example of such processing is image data compression processing. Specifically, the control unit 11 stores the image data after the background removal processing in the storage unit 12, and performs the next compression processing on the processing target. As shown functionally in FIG. 5, this compression processing includes a pattern candidate portion specifying processing unit 41 that specifies a portion (design candidate portion) that is a candidate pattern portion for the image data to be processed, A character line drawing extraction processing unit 42 that extracts line drawings, a layout processing unit 43 that performs layout processing, a same color character region separation unit 44, a hole filling processing unit 45, and a compression processing unit 46 are configured. .

  The pattern candidate portion specifying processing unit 41 binarizes the image data to be processed, for example, with a predetermined threshold value regarding brightness (“white” is a portion higher than the brightness threshold value, and is lower than the brightness threshold value) "Black"), demarcate the connected pixel area of black pixels in the binarized image data, and select an area of the connected pixel area larger than a predetermined size threshold as a pattern candidate part Identify.

  The character / line drawing extraction processing unit 42 specifies an image region having a property as a character / line drawing as a character / line drawing region from a processing target image data stored in the storage unit 12 by a widely known method.

  The layout processing unit 43 performs a layout process on the image specified as the pattern candidate part, removes a part in which the arrangement state of the plurality of connected pixels is recognized as the character string arrangement from the pattern candidate part, Include in Further, the pattern candidate portion after this processing (that is, the portion from which the arrangement as a character string has been removed as a result of the layout processing) is specified as the pattern portion.

  The same color character area separation unit 44 generates image data for each color for the character / line drawing in the area specified as the character line drawing area among the processing target image data stored in the storage unit 12. For each piece of image data, information specifying the color and data obtained by binarizing the image data and run-length compression are generated, and the character / line drawing area compressed data is generated by associating them. At this time, each character / line drawing compressed data is associated with coordinate information for specifying a portion from which the character / line drawing portion is extracted.

  On the other hand, the hole-filling processing unit 45 extracts image data of a part specified as a pattern part from the image data to be processed stored in the storage unit 12, and includes pixels (characters / characters) not included in the part. The pixel of the portion extracted as a line drawing) is determined and set (filled) by a predetermined method.

  The compression processing unit 46 compresses the image data of the pattern portion after the hole filling process output from the hole filling processing unit 45 by a compression method suitable for pattern compression such as JPEG (Joint Picture Experts Group). The line drawing area compressed data is concatenated and written to the storage unit 12 as data in a predetermined format (for example, PDF (Portable Document Format) data).

  By the processing of the control unit 11, it becomes possible to compress the character / line drawing and the picture by a compression process suitable for each of them and generate data with improved compression efficiency.

  When decompressing this data, the compression processing result of the pattern portion and the character / line drawing area compression data are separated, and the decompression processing corresponding to the compression processing (for example, JPEG decompression processing) is performed on the pattern portion. .

  On the other hand, for character / line drawing parts, run-length compression is expanded to obtain binarized image data and color information to be attached thereto, and the black pixel part of the binarized image data is replaced with the color information. The image data for each color is restored by setting the pixel value indicated by. Then, the image data for each color of each character / line drawing part is opaquely combined with the corresponding position (position specified by the associated coordinate information) of the decompressed picture part to restore the original image data.

1 is a configuration block diagram of an image processing apparatus according to an embodiment of the present invention. It is a functional block diagram showing the process example of the image processing apparatus which concerns on the 1st Embodiment of this invention. It is explanatory drawing showing the example of saturation correction | amendment LUT. It is a functional block diagram showing the process example of the image processing apparatus which concerns on the 2nd Embodiment of this invention. It is a functional block diagram showing an example of the process which is a usage example with respect to the image data after correction | amendment. It is a functional block diagram of the conventional image processing apparatus. It is explanatory drawing showing the example of the brightness correction | amendment LUT. It is explanatory drawing showing the example of the brightness histogram before and behind correction | amendment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Lightness histogram calculation part, 2 Background threshold determination part, 3 LUT production | generation part, 4 Lightness correction part, 11 Control part, 12 Storage part, 13 Image input part, 14 Image output part, 21 Color space conversion part, 22 Brightness feature-value Calculation unit 23 Brightness background threshold determination unit 24 Background region extraction unit 25 Brightness correction LUT generation unit 26 Saturation feature amount calculation unit 27 Saturation correction LUT generation unit 28 Correction unit 31 Distance plane calculation unit 32 Distance feature amount calculation unit, 33 correction parameter generation unit, 41 picture candidate part specification processing unit, 42 character line drawing extraction processing unit, 43 layout processing unit, 44 same color character region separation unit, 45 burial processing unit, 46 compression processing unit .

Claims (7)

An image processing apparatus that performs processing to remove a background portion that is a background of a significant image from image data to be processed,
Means for extracting a background area from the image data to be processed;
Means for generating a color correction parameter based on color relationship information related to a color component other than lightness represented by a pixel in the extracted background region;
Correction means for correcting the color of the image data to be processed using the generated color correction parameter;
An image processing apparatus comprising: The image processing apparatus according to claim 1,
The color-related information is saturation information represented by pixels in the background area,
The color correction parameter is a parameter for correcting saturation, and the means for correcting the color corrects the saturation of each pixel included in the image data to be processed using the color correction parameter. An image processing apparatus. The image processing apparatus according to claim 2,
Means for generating a parameter for correcting the brightness based on the information related to the brightness represented by the pixel in the extracted background region;
The image processing apparatus according to claim 1, wherein the correction unit corrects a brightness component of the image data to be processed using a parameter for correcting the brightness and the color correction parameter. The image processing apparatus according to claim 1,
The image processing apparatus according to claim 1, wherein the color relationship information is information relating to a distance between a pixel value in a base area in a color space and a specific pixel value. The image processing apparatus according to claim 4,
The information regarding the distance is a distance defined by assigning different weights to each of the brightness component and the other components. An image processing method for performing a process of removing a background portion that becomes a background of a significant image from image data to be processed using a computer,
Extracting a background area from the image data to be processed;
Generating color correction parameters based on color relationship information related to color components other than lightness represented by the pixels in the extracted background region;
Correcting the color of the image data to be processed using the generated color correction parameter;
An image processing method comprising: An image processing program for causing a computer to perform a process of removing a background portion that is a background of a significant image from image data to be processed,
A procedure for extracting a background area from the image data to be processed;
A procedure for generating a color correction parameter based on color relationship information related to a color component other than lightness represented by a pixel in the extracted background region;
A procedure for correcting the color of the image data to be processed using the generated color correction parameter;
An image processing program for causing a computer to execute.

Images