JP2006050559A - Image correcting apparatus and method, and image correction program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To perform proper image correction processing using a plurality of color correction values. <P>SOLUTION: Provided image data are used, to respectively calculate two color correction values in two correction value calculating circuits 12, 13. Further, a target skin color chromaticity value and significance levels indicating which of the two color correction values should be regarded as important are inputted from an input device 2. Respective reliability for the two color correction values are calculated on the basis of the two color correction values and the inputted target skin color chromaticity value. On the basis of the calculated reliability, the inputted importance and the two color correction values, both of the reliability and the importance are reflected, to calculate a new color correction value in a color correction value calculating circuit 15. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus and method for performing color correction processing (color conversion processing) on digital image data, and a program for causing a computer to execute the image processing method.

  When creating a color print from digital image data obtained by reading an image recorded on a photographic negative film or reversal film with a scanner, or from digital image data acquired by a digital camera or camera-equipped mobile phone, The digital image data is subjected to a color correction process for correcting the color balance so as to create an image similar to the scene at the time of shooting, thereby creating a print image. Estimate the light source at the time of shooting or estimate the color balance of the entire image from the image acquired by shooting, and perform color correction processing so that the achromatic color becomes an appropriate achromatic color. We try to be appropriate.

  In Patent Document 1, a face area is first detected from input image data, and color conversion processing (color correction) is performed on the entire image data so that the color of the detected face area becomes a desired color. Yes.

In Patent Document 1, color correction is performed by paying attention only to a face area (skin color area) in an image. Since face areas (skin color areas) vary among individuals, there is a high possibility that color correction will be unnatural.
JP 2000-182043 A

  The present invention has been made in view of the above circumstances, and an object thereof is to enable more appropriate color correction (color balance correction) of image data.

  Another object of the present invention is to enable color correction of image data in accordance with the user's wishes.

  An image correction apparatus according to the present invention is calculated by a plurality of types of color correction value calculation means for calculating color correction values by different color correction value calculation processes based on given image data, and the plurality of types of color correction value calculation means. Reliability calculation means for calculating reliability for each of the plurality of color correction values obtained, and using the reliability for each of the plurality of color correction values calculated by the reliability calculation means, the plurality of color corrections Each of the values is weighted, and an added color correction value calculating means for calculating a new color correction value obtained by adding a plurality of weighted color correction values, and the new color calculated by the added color correction value calculating means And a correction means for correcting the color of the image data based on a correct color correction value.

  The image correction method according to the present invention calculates color correction values by different color correction value calculation processes based on given image data, calculates the reliability for each of the calculated color correction values, and calculates Using the reliability for each of the plurality of color correction values, a new color correction value obtained by weighting each of the plurality of color correction values and adding the plurality of weighted color correction values is calculated. The image data is color-corrected based on the calculated new color correction value.

  The present invention further provides a program for causing a computer to execute image correction processing (a program for causing a computer to function as an image correction apparatus). This program is based on image data given from a storage device (hard disk, memory card, CD-ROM, DVD-ROM, etc.), and each color correction value is calculated by different color correction value calculation processing based on the given image data. Weighting each of the plurality of color correction values using the processing for calculating the reliability, the processing for calculating the reliability for each of the plurality of calculated color correction values, and the reliability for each of the plurality of calculated color correction values. Then, a process for calculating a new color correction value obtained by adding a plurality of weighted color correction values, and a process for correcting the color of the image data based on the calculated new color correction value are performed on the computer. To be executed. All of these processes may be executed by a CPU included in the computer, or a part of the processes may be executed by a dedicated hardware device.

  Based on the given image data, color correction values are calculated by different color correction value calculation processes. That is, a plurality of color correction values are calculated.

  Multiple types of color correction value calculation processing are common in terms of color correction value calculation processing, but the processing content (data, algorithm, etc. used for color correction value calculation) is different. Each of the calculation processes is meant. For example, a process of calculating a color correction value so that color information representing one specific color (for example, skin color, white color, etc.) included in an image represented by image data has color information that conforms to ideal color information, Color correction value is calculated so that the color information representing a specific color (for example, skin color and gray color) has color information that matches the ideal color information, the color temperature and the light source type are estimated, and the estimated color temperature and light source A process for calculating a color correction value corresponding to a seed is known. Of these multiple types of color correction value calculation processes, any two or more calculation processes can be employed as the color correction value calculation process (means).

  The reliability is calculated for each of the plurality of color correction values calculated by the plurality of types of color correction value calculation means. This reliability indicates whether or not the calculated color correction value is a value that allows appropriate color correction (the degree thereof).

  For example, the color information about the representative color included in the image data, the ideal color information about the representative color, and the plurality of color correction values calculated by the plurality of types of color correction value calculation means, Based on the plurality of corrected color information obtained by correcting the color information, the reliability for each of the plurality of color corrections is calculated.

  For example, in a color space (for example, chromaticity space), an ideal line segment vector obtained by connecting the color information and the ideal color information, the color information, and the plurality of corrected color information, respectively. The angle formed by each of the plurality of correction line segment vectors obtained by connecting the two can be treated as a value indicating how reliable each of the plurality of color correction values is. If the angle formed between the ideal line segment vector and a correction line segment vector based on a certain color correction value is small, it can be evaluated that the color correction value is highly likely to realize ideal color correction. In this case, the reliability with respect to the color correction value becomes high (a reliability with a large value is associated). Conversely, if the angle between the ideal line vector and a correction line vector based on a certain color correction value is large, the color correction value may be a color correction value that performs color correction far from the ideal. Is expensive. In this case, the reliability of the color correction value is low (the reliability with a small value is associated).

  Of course, reliability calculation other than the reliability calculation based on the angle formed by the ideal line segment vector and the correction line segment vector may be employed. In the color space, the reliability of each of the plurality of color correction values is determined based on the size of each of the plurality of correction line segment vectors obtained by connecting each of the color information and the plurality of corrected color information. In the color space, a plurality of color correction values for each of a plurality of color correction values may be calculated according to the directions of a plurality of correction line segment vectors obtained by connecting each of the color information and the plurality of corrected color information. The reliability may be calculated. A value obtained by calculating a plurality of reliability levels for one color correction value and multiplying the reliability levels may be treated as the reliability level of the color correction value.

  The reliability for each of the plurality of calculated color correction values is used, each of the plurality of calculated color correction values is weighted, and a new color correction is performed by adding the plurality of weighted color correction values. A value is calculated. The image data is color-corrected using the calculated new color correction value.

  According to the present invention, a new color correction value used for color correction of image data is obtained by weighting each of the plurality of color correction values according to reliability and adding the plurality of weighted color correction values. Therefore, the value (color correction value) reflects the reliability of the plurality of color correction values. Since the possibility that an inappropriate color correction value is calculated is reduced, a more appropriate color correction value can be obtained.

  Preferably, the importance for each of the plurality of color correction values is input. This input of importance is intended to incorporate the user's wishes, and among the plurality of color correction values, color correction that the importance of a large value is not important for the color correction value that the user considers important. For the value, the importance of a small value is input. In this case, the reliability for each of the plurality of calculated color correction values and the importance for each of the plurality of input color correction values are used, and each of the plurality of calculated color correction values is weighted. It is done. A more appropriate color correction value and a color correction value according to the user's preference can be obtained.

  Of course, either reliability or importance may be weighted more heavily in weighting. It is possible to obtain a new color correction value strongly influenced by the reliability and a new color correction value strongly influenced by the importance.

  Embodiments of the present invention will be described below with reference to the accompanying drawings.

  FIG. 1 is a block diagram showing the overall configuration of the digital print system. The digital print system includes an image correction device 1 and peripheral devices (input device 2, display device 3, storage device 4, and printer 5) connected to the image correction device 1. FIG. 2 is a block diagram showing a detailed electrical configuration of the image correction apparatus 1 which is a core apparatus of the digital printer system together with a data flow. In FIG. 2, the display device 3 and the printer 5 are omitted from the peripheral devices shown in FIG.

  The image correction apparatus 1 is an apparatus for correcting image data so that the color balance of an image represented by given image data has a more appropriate color balance. The image correction apparatus 1 performs color balance correction on given image data and outputs the image data after color balance correction. Hereinafter, in this specification, the color balance correction is simply referred to as “color correction”.

  An input device 2 (keyboard, mouse, etc.) connected to the image correction device 1 is used to input parameters for color correction processing (target (ideal) skin color chromaticity value and two importance levels). On the display screen of the display device 3 (CRT display, liquid crystal display, etc.), a screen for setting parameters input from the input device 2, an image represented by image data before and after color correction, and the like are displayed. Is done. Image data is stored in a storage device (hard disk, memory card, CD-ROM, etc.) 4, and color correction processing is performed in the image correction device 1 on the image data read from the storage device 4. . The printer 5 prints an image represented by the color-corrected image data on photographic paper or the like.

  Referring to FIG. 2, an image correction apparatus 1 includes a face area partition circuit 11, a color correction value calculation circuit based on a skin color area (first color correction value calculation circuit) 12, and a color correction value calculation circuit based on a gray color area. (Second color correction value calculation circuit) 13, a reliability calculation circuit 14, a color correction value calculation circuit 15 based on importance and reliability, and an image correction circuit 16 are provided.

  In the image correction apparatus 1, the image correction circuit 16 performs color correction processing on image data read from the storage device 4 (hereinafter referred to as original image data). Details of the color correction processing performed in the image correction circuit 16 will be described later.

  The color correction processing performed in the image correction circuit 16 follows the color correction value calculated in the color correction value calculation circuit 15 based on importance and reliability. The color correction value calculation circuit 15 based on the importance and the reliability includes a reliability calculated in the reliability calculation circuit 14 (details will be described later), an importance input from the input device 2 (details will be described later), and skin color. Based on the color correction value obtained by the color correction value calculation circuit 12 based on the area and the color correction value obtained by the color correction value calculation circuit 13 based on the gray color area, it is used for the color correction processing performed in the image correction circuit 16. A color correction value (new color correction value) to be calculated is calculated.

  First, a color correction value calculation process based on a skin color area and a color correction value calculation process based on a gray color area will be described. In the color correction value calculation process based on the skin color area and the color correction value calculation process based on the gray color area, the face area partition circuit 11, the color correction value calculation circuit 12 based on the skin color area, and the color correction value based on the gray color area The calculation circuit 13 operates.

  The image represented by the original image data includes an image area representing a human face (hereinafter referred to as a face image portion) and a gray image area (an image area representing a gray object or the like; hereinafter a gray image). Part)). Original image data representing an image including a face image portion and a gray image portion is read from the storage device 4.

  The original image data read from the storage device 4 is input to the face area dividing circuit 11, the color correction value calculation circuit 12 based on the skin color area, and the color correction value calculation circuit 13 based on the gray color area.

  The face area partitioning circuit 11 performs processing for partitioning a face image portion included in an image represented by original image data (hereinafter referred to as an original image). For the section processing of the face image portion performed in the face area partitioning circuit 11, various conventional or new section (detection) techniques can be used. For example, the image is divided into a plurality of areas, photometry is performed, the data obtained by photometry is converted into hue and saturation, these two-dimensional histograms are created, and this histogram is decomposed into unimodal peaks, Further, it is judged which pixel each pixel of the image belongs to, and the pixels are divided into groups corresponding to the divided mountains, and the face candidate region is extracted by dividing the image for each group, and the extracted region By comparing the image size with the image size obtained from the main image size information, an extraction method can be adopted in which the most matching region is used as the face region (see Japanese Patent Laid-Open No. 5-100328). After dividing the image into multiple regions, excluding non-human regions, a shape pattern representing the contour of the person's head is detected, a face candidate region is detected according to the detected pattern, and the contour of the person's face is represented The shape pattern, the shape pattern representing the internal structure of the face, and the shape pattern representing the contour of the human torso are detected, and the consistency of the face candidate area is determined based on the detected pattern, and the accuracy corresponding to the human face A method of extracting a face region by determining a region having the highest value (see Japanese Patent Laid-Open No. 8-122944) may be used. A learning frame is set in the image, image data representing an image portion defined by the learning frame is learned, one or more representative feature vectors representing the learning frame are extracted, and the search is performed in the image In addition to setting a frame, a subject area is extracted from the search frame based on the similarity between each of the above representative feature vectors and a plurality of feature vectors extracted from image data representing the image portion defined by the search frame. A technique (Japanese Patent Application No. 2002-378989) can also be used.

  Of course, the operator of the digital print system may use the input device 2 (such as a mouse) to partition the face image portion in the original image from other image portions (manual face region partition). In this case, the face area dividing circuit 11 is not necessarily required for the image correction apparatus 1.

  The face area dividing circuit 11 includes 1 (or 0) of pixels included in the sectioned (detected) face image part (in the face image area) and pixels included in an image part other than the face image part (outside the face image area). Binary data (data in which 1 or 0 is associated with each pixel address) with 0 (or 1) being output. 3A shows an example of an original image, FIG. 3B shows a state in which the face image portion is divided by the face area dividing circuit 11, and FIG. 3C shows each pixel included in the face image portion. 1 schematically shows binary data (mask data) output from the face area dividing circuit 11 in which each pixel included in an image portion other than the face image portion is set to 0.

  The binary data output from the face area dividing circuit 11 indicates the position and range (area position) of the face image portion included in the original image. Hereinafter, the binary data output from the face area dividing circuit 11 is referred to as face area information.

  The face area information output from the face area dividing circuit 11 is input to the color correction value calculation circuit 12 based on the skin color area and the color correction value calculation circuit 13 based on the gray color area, respectively. Further, the target skin color chromaticity value (ideal chromaticity value appropriate as the chromaticity representing the skin color) input from the input device 2 is also based on the color correction value calculation circuit 12 based on the skin color area and the gray color area. Each is input to the color correction value calculation circuit 13.

  The color correction value calculation circuit 12 based on the skin color region is configured so that the color (skin color) of the face image portion (which is a portion that surely includes many skin color pixels) has a chromaticity value that matches the target skin color chromaticity value. This is a circuit for calculating a color correction value (coefficient) for color correction. The color correction value calculation circuit 13 based on the gray color region is a circuit that calculates a color correction value (coefficient) for performing color correction so that the gray image portion has a chromaticity value along the target gray chromaticity value. Although the color correction value calculation circuit 12 based on the skin color area and the color correction value calculation circuit 13 based on the gray color area are different in the image portion, color correction value calculation processing, algorithm, and the like used for calculating the color correction value, they are all This is common in that it is a circuit for calculating a color correction value.

  The color correction value calculation circuit 12 based on the skin color area extracts the data representing the face image part included in the original image data using the face area information given from the face area dividing circuit 11, and represents the extracted face image part. A representative RGB value (for example, an average RGB value) obtained from the RGB value for each pixel constituting the face image portion represented by the data is calculated. Then, based on the average skin color chromaticity value obtained by converting the average RGB value into the chromaticity value and the target skin color chromaticity value input from the input device 2, a gain coefficient for the R value, a gain coefficient for the G value, And a gain coefficient for the B value is calculated. For example, these gain coefficients can be obtained by a method described in Japanese Patent Application Laid-Open No. 2000-182043 (calculation of gain coefficients using chromaticity values).

The gain coefficient for the R value obtained by the color correction value calculation circuit 12 based on the skin color region is G _r1 , the gain coefficient for the G value is G _g1 , and the gain coefficient G _b1 for the B value (collectively, the first The following input / output relationship of Equation 1 is obtained.

R ₁ = G _r1 · R
G ₁ = G _g1 · G Equation 1
B ₁ = G _b1・ B

In Equation 1, R, G and B are R values, G values and B values for each pixel constituting the original image, and R ₁ , G ₁ and B ₁ are obtained by the color correction value calculation circuit 12 based on the skin color area. R values, G values, and B values after color correction when color correction processing based on the gain coefficient is performed.

Note that the color correction value calculation circuit 12 based on the skin color area calculates G _r1 , G _g1, and G _b1 using the above gain coefficients, and does not perform image correction itself. For convenience of explanation, the above formula 1 is expressed by the following formula 2.

Color correction value based on skin color area (first gain coefficient) = (G _r1 , G _g1 , G _b1 )

  Note that the first gain coefficient can be obtained by using the chromaticity value, or the first gain coefficient can be obtained based on the RGB value or the true RGB value. Further, the first gain coefficient may be further adjusted so that the brightness of the image does not change due to the color correction.

  For example, the color correction value calculation circuit 13 based on the gray color region calculates the gain coefficient corresponding to the R value, the gain coefficient corresponding to the G value, and the gain coefficient corresponding to the B value based on the method described in Japanese Patent Laid-Open No. 2003-209856. Can be calculated. The gain coefficient can be calculated as follows.

(i) A representative RGB value (for example, an average RGB value) of a face image portion (skin color region) is obtained, and the obtained average RGB value is converted into a chromaticity value (average skin color chromaticity value) and obtained by conversion. An average skin color temperature corresponding to the average skin color chromaticity value is obtained.
(ii) A target skin color color temperature corresponding to the target skin color chromaticity value input from the input device 2 is obtained.
(iii) A gray color temperature (target gray color temperature) corresponding to the target skin color temperature is obtained, and a target gray color chromaticity value corresponding to the obtained target gray color temperature is obtained.
(iv) A gray color region (gray color pixel group) in the original image is detected using at least the average skin color temperature of the face image portion.
(v) A chromaticity value representative of the detected gray color area (for example, an average gray color chromaticity value) is obtained.
(vi) The average skin color temperature and the average gray color chromaticity value are corrected so that the number of pixels in the gray color region (gray pixel group) is maximized ((iv) to (vi) are repeated).
(vii) Based on the target gray color chromaticity value, the corrected average gray color chromaticity value, and the average gray color chromaticity value before correction, the gain coefficient G _r2 for the R value, the gain coefficient G _g2 for the G value, and B A gain coefficient G _b2 for the value is calculated.

Using the gain coefficients G _r2 , G _g2 , and G _b2 (collectively referred to as the second gain coefficient) obtained by the color correction value calculation circuit 13 based on the gray color region, Output relationship is obtained.

R ₂ = G _r2 · R
G ₂ = G _g2 · G Equation 3
B ₂ = G _b2・ B

In Equation 3, R, G, and B are R values, G values, and B values for each pixel constituting the original image, and R ₂ , G ₂ , and B ₂ are obtained by the color correction value calculation circuit 13 based on the gray color region. R values, G values, and B values after color correction when color correction processing based on the obtained gain coefficient is performed.

Note that the color correction value calculation circuit 13 based on the gray color area calculates the above-described second gain coefficients G _r2 , G _g2 , and G _b2 , and does not perform image correction itself. For convenience of explanation, the above expression 3 is expressed by the following expression 4.

Color correction value based on gray color region (second gain coefficient) = (G _r2 , G _g2 , G _b2 )
..Formula 4

  The color correction value (first gain coefficient) calculated by the color correction value calculation circuit 12 based on the skin color area and the color correction value (second gain coefficient) calculated by the color correction value calculation circuit 13 based on the gray color area Are given to the reliability calculation circuit 14.

  Next, the process of the reliability calculation circuit 14 will be described.

  The reliability calculation circuit 14 includes original image data, face area information output from the face area dividing circuit 11, a color correction value (first gain coefficient) based on a skin color area, and a color correction value (first gain coefficient based on a gray color area). 2) and the target skin color chromaticity value input from the input device 2 are input.

  The reliability calculation circuit 14 calculates the color correction value (first gain coefficient) calculated in the color correction value calculation circuit 12 based on the above-described skin color area and the color correction value calculation circuit 13 based on the gray color area. When the original image data is color corrected using the color correction value (second gain coefficient), determine how far the corrected image color is from the target image color. A numerical value given according to the degree of deviation (this numerical value is referred to as “reliability”) is obtained.

  Several methods can be employed to calculate the reliability (which may be called accuracy or appropriateness). Here, representative reliability calculation will be described in detail.

  In typical reliability calculation, a chromaticity space (rb space) is used. 4 and 5 show a chromaticity space (rb space).

The following three points are plotted in the chromaticity space shown in FIG.
(1) Representative skin color chromaticity values (ro, bo) based on data representing the face image portion (skin color region) in the original image (for example, the average chromaticity value of the face image portion is used as the representative skin color chromaticity value)
(2) Target skin color chromaticity value (rt, bt) (input from the input device 2)
(3) Representative skin color chromaticity value (r1, b1) after color correction according to the color correction value (first gain coefficient) based on the skin color area

  In the chromaticity space shown in FIG. 5, the following three points are plotted.

(1) Representative skin color chromaticity values (ro, bo) based on data representing the face image part (skin color region) in the original image
(2) Target skin color chromaticity value (rt, bt)
(3) Representative skin color chromaticity value (r2, b2) after color correction according to the color correction value (second gain coefficient) based on the gray color region

  The chromaticity space shown in FIG. 4 is ideally color-corrected (line segment vector A) so that the representative skin color chromaticity values (ro, bo) have the target skin color chromaticity values (rt, bt). When the color correction based on the above equation 1 is performed, the representative skin color chromaticity value (ro, bo) is color corrected to the representative skin color chromaticity value (r1, b1) after color correction (line segment vector B1 ). In the chromaticity space shown in FIG. 4, the angle θ1 formed by the line segment vector A and the line segment vector B1 is corrected using the color correction value (gain coefficient) obtained by the correction value calculation circuit 12 based on the skin color area. It is treated as the degree of deviation from the ideal color correction.

  Similarly, in the chromaticity space shown in FIG. 5, color correction (line segment vector A) is performed so that the representative skin color chromaticity values (ro, bo) have the target skin color chromaticity values (rt, bt). Ideally, if color correction based on Equation 3 (Equation 4) is performed, the representative skin color chromaticity value (ro, bo) is corrected to the representative skin color chromaticity value (r2, b2) after color correction. (Line segment vector B2). In the chromaticity space shown in FIG. 5, the angle θ2 formed by the line segment vector A and the line segment vector B1 is determined using the color correction value (gain coefficient) obtained by the correction value calculation circuit 13 based on the gray color area. It is treated as the degree of deviation from the ideal color correction when corrected.

  FIG. 6 shows an example of a function graph showing the relationship between the angle θt (t = 1, 2,...) Representing the above-described divergence and the reliability ft (θt) calculated according to this angle. ing. The reliability is obtained as any value from 0 to 1.

  Although the representative reliability calculation has been described above, the magnitudes of the line segment vector B1 and the line segment vector B2 may be used as the reliability, and the direction of the line segment vector B1 itself and the direction of the line segment vector B2 itself are determined as reliability. It is good. In the chromaticity space, the reliability may be calculated based on the degree of proximity of the skin color representative color to the skin black body locus and the degree of proximity of the gray color representative color to the gray black body locus. A value obtained by multiplying any or all of the plurality of reliability levels may be set as the reliability level. In any case, the reliability calculation circuit 14 includes the color correction value (first gain coefficient) calculated in the color correction value calculation circuit 12 based on the above-described skin color area and the color correction value calculation circuit based on the gray color area. With respect to the color correction value (second gain coefficient) calculated in step 13, the reliability is calculated according to how far the corrected image color deviates from the target image color.

  The reliability of the color correction value based on the skin color area (hereinafter referred to as skin color reliability) and the reliability of the color correction value based on the gray color area (hereinafter referred to as gray color reliability) obtained by the reliability calculation circuit 14 are as follows. The color correction value calculation circuit 15 based on importance and reliability is input.

  Next, processing of the color correction value calculation circuit 15 based on importance and reliability will be described. The color correction value calculation circuit 15 based on the importance and reliability inputs the above-described skin color reliability, gray color reliability, and importance input from the input device 2. The importance level refers to the color correction value (first gain coefficient) calculated by the color correction value calculation circuit 12 based on the skin color area and the color correction value (first gain coefficient) calculated by the color correction value calculation circuit 13 based on the gray color area ( Of the second gain coefficient), this is a value representing how much color correction value is used with emphasis, and reflects the user's preference (request).

  The color correction value calculation circuit 15 based on importance and reliability executes any one of the following three calculation processes.

(First calculation process)
In the first calculation process, a color correction value (Gr, Gg, Gb) (hereinafter referred to as “new color correction value”) based on importance and reliability is calculated by the following equation (5).

  In Equation 5, w t is expressed by Equation 6 and Equation 7 below.

  In Expressions 5 to 7, the variable t distinguishes the type of color correction value (or color correction value calculation circuit). In this embodiment, since two types of color correction values are calculated, t = 1 or 2. The variable k is the number of calculated color correction values. In this embodiment, k = 2. In addition, w t (in this embodiment, w 1 and w 2) in Expression 5 to Expression 7 represents a weight for each calculated color correction value. In the equation 6, It (I1 and I2 in this embodiment) is the importance for each color correction value input from the input device 2, and Ct (C1 and C2 in this embodiment) is the color. The reliability for each correction value is shown. These are the same in the second calculation process and the third calculation process described later.

  In the first calculation process (correction value calculation based on Expressions 5 to 7), a color correction value having a higher reliability is greatly reflected by a new color correction value (Gr, Gg, Gb) and is larger. It can be seen that the color correction value having importance is greatly reflected by the new color correction values (Gr, Gg, Gb). A new color correction value reflecting a more appropriate color correction value can be obtained while depending on which color correction value the user places importance on (the user's preference).

  Needless to say, when a new color correction value is calculated based only on the reliability, the two importance levels input from the input device 2 may be set to the same value (for example, I1 = 1, I2 = 1). Yes.

(Second calculation process)
In the second calculation process, a new color correction value (Gr, Gg, Gb) is calculated by the following equation 8.

  In Expression 8, the weight w t is expressed by Expression 9 shown below.

  In the second calculation process (Equation 8 and Equation 9), the importance It is reflected more greatly, and a new color correction value is obtained. Note that the value obtained by adding all the weights w t is “1” (the above formula 7) is the same as the first calculation process.

(Third calculation process)
The third calculation process is the same as the second calculation process described above in that a new color correction value (Gr, Gg, Gb) is calculated by the above-described equation 8. The third calculation process is different in that the weight wt is expressed by the following formula 10.

  In the third calculation process (Equation 8 and Equation 10), a new color correction value in which the reliability Ct is more largely reflected is obtained. Note that the value obtained by adding all the weights w t is “1” (the above formula 7) is the same as the first calculation process.

  Which one of the first to third calculation processes described above should be used may be selected by the user using the input device 2, for example. A new color correction value (Gr, Gg, Gb) is calculated by any one of the selected first to third calculation processes.

  The new color correction values (Gr, Gg, Gb) calculated by the color correction value calculation circuit 15 based on importance and reliability are input to the image correction circuit 16. Based on the new color correction values (Gr, Gg, Gb), the original image data (which may be only data representing the face image portion) is color-corrected.

  The color correction value (first gain coefficient) calculated in the color correction value calculation circuit 12 based on the above-described skin color area and the color correction value (first gain coefficient) calculated in the color correction value calculation circuit 13 based on the gray color area. 2) is completely different (for example, when the direction of the above-described line segment vector B1 and line segment vector B2 is completely different), the calculation of one of the color correction values has failed. It is highly possible that In such a case, the image correction circuit 16 may weaken the color correction effect by the new color correction values (Gr, Gg, Gb).

  In the embodiment described above, based on the two color correction values of the color correction value calculated by the color correction value calculation circuit 12 based on the skin color area and the color correction value calculated by the color correction value calculation circuit 13 based on the gray color area. Although a new color correction value is calculated, a color correction value calculated by another method (for example, a color correction calculated based on a color correction value calculation process described in Japanese Patent Application Laid-Open No. 2000-148978). Value) may be used instead of or in addition to one of the two color correction values calculated in this embodiment.

  In the above-described embodiment, the weight and reliability are calculated and the importance is input for each color correction value calculation circuit (color correction value). May be calculated and the degree of importance may be input.

  In the above-described embodiments, the gain coefficient is used as the color correction value. However, it goes without saying that a lookup table storing the correspondence between the input value and the output value may be used as the color correction value instead of the gain coefficient. Yes.

1 is a block diagram illustrating an overall configuration of a digital print system. It is a block diagram which shows the detailed electrical structure of an image correction apparatus with the flow of data. (A) shows an example of an image represented by original image data, (B) shows a state in which a face image portion is partitioned (detected), and (C) shows an example of a mask image. An angle formed by a line segment vector connecting the representative skin color chromaticity value and the target skin color chromaticity value and a line segment vector connecting the representative skin color chromaticity value and the corrected representative skin color chromaticity value according to one color correction value Show. The angle formed by the line segment vector connecting the representative skin color chromaticity value and the target skin color chromaticity value and the line segment vector connecting the representative skin color chromaticity value and the representative skin color chromaticity value after correction according to other color correction values Indicates. The function graph for calculation of reliability is shown.

Explanation of symbols

1 Image Correction Device 2 Input Device 3 Display Device 4 Storage Device 5 Printer
11 Face area division circuit
12 Color correction value calculation circuit based on skin color area
13 Color correction value calculation circuit based on gray color area
14 Reliability calculation circuit
15 Color correction value calculation circuit based on importance and reliability
16 Image correction circuit

Claims (10)

A plurality of types of color correction value calculation means for calculating color correction values by different color correction value calculation processes based on given image data;
Reliability calculation means for calculating reliability for each of the plurality of color correction values calculated by the plurality of types of color correction value calculation means;
Using the reliability for each of the plurality of color correction values calculated by the reliability calculation means, each of the plurality of color correction values is weighted, and the weighted color correction values are added. An added color correction value calculating means for calculating a new color correction value, and a correcting means for color correcting the image data based on the new color correction value calculated by the added color correction value calculating means,
An image correction apparatus comprising: An importance input means for inputting importance for each of the plurality of color correction values calculated by the plurality of types of color correction value calculation means;
The additional color correction value calculation means is:
Using the reliability for each of the plurality of color correction values calculated by the reliability calculation means and the importance for each of the plurality of color correction values input from the importance input means, the plurality of color corrections Each of the values is weighted,
The image correction apparatus according to claim 1. The reliability calculation means is:
The color information based on each of color information about the representative color included in the image data, ideal color information about the representative color, and a plurality of color correction values calculated by the plurality of types of color correction value calculation means. The image correction apparatus according to claim 1, wherein the reliability of each of the plurality of color corrections is calculated based on the plurality of corrected color information obtained by correcting the color. The reliability calculation means is:
In a color space, an ideal line segment vector obtained by connecting the color information and the ideal color information, and a plurality of correction line segment vectors obtained by connecting each of the color information and the plurality of corrected color information. The image correction apparatus according to claim 3, wherein the reliability of each of the plurality of color correction values is calculated in accordance with an angle formed by each of the image correction apparatus. The reliability calculation means is:
In the color space, the reliability of each of the plurality of color correction values is determined based on the size of each of the plurality of correction line segment vectors obtained by connecting each of the color information and the plurality of corrected color information. The image correction apparatus according to claim 3, which is to be calculated. The reliability calculation means is:
Calculating the reliability of each of a plurality of color correction values according to the directions of a plurality of correction line segment vectors obtained by connecting each of the color information and the plurality of corrected color information in a color space; The image correction apparatus according to claim 3, wherein Based on the given image data, each color correction value is calculated by different color correction value calculation processing,
Calculate the reliability for each of the calculated color correction values,
A new color correction value obtained by weighting each of the plurality of color correction values and adding the plurality of weighted color correction values using the reliability of the calculated plurality of color correction values. Calculate
Based on the calculated new color correction value, color correction is performed on the image data.
Image correction method. Accept the input of importance for each of the calculated color correction values,
Each of the plurality of color correction values is weighted using the reliability for each of the calculated plurality of color correction values and the importance for each of the plurality of input color correction values.
The image correction method according to claim 7. A process for calculating color correction values by different color correction value calculation processes based on given image data,
A process for calculating the reliability of each of the calculated color correction values,
A new color correction value obtained by weighting each of the plurality of color correction values and adding the plurality of weighted color correction values using the reliability of the calculated plurality of color correction values. A process for calculating and color correcting the image data based on the calculated new color correction value;
Correction program that causes a computer to execute Causing the computer to further execute a process of accepting an input of importance for each of the calculated color correction values,
The process of calculating the new color correction value is as follows:
The weight of each of the plurality of color correction values is weighted using the reliability for each of the calculated plurality of color correction values and the importance for each of the plurality of input color correction values. Image correction program.

Images