JP5633172B2 - Image processing apparatus and image processing program - Google Patents

Description

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

  When a person is photographed using a flash, a red-eye phenomenon occurs in which the eye color is photographed in red or gold. Some apparatuses having an image processing function such as a digital camera, a computer, or a printer can correct an image in which the red-eye phenomenon occurs by image processing (see Patent Document 1).

JP-A-10-233929

  However, in the conventional method, the region where the red-eye phenomenon occurs cannot be accurately identified, and as a result, there are cases where the red-eye correction cannot be performed appropriately.

  The present invention has been made in view of such circumstances, and it is an object of the present invention to provide an image processing apparatus and an image processing program that can appropriately specify a region where a red-eye phenomenon occurs.

To achieve the above object, a first According to one aspect, an image processing apparatus for identifying redeye pixels from the red-eye candidate region including a red eye that occurrence of the red-eye effect in the color image of the present invention, red Based on the value of the color element of the pixel group in the red-eye candidate region in a three-axis space in which at least one color element is a redness value with a color element as an axis for a color image specifying a pixel An ellipsoid defining means for defining an ellipsoid that separates the red-eye pixel and the non-red-eye pixel, and, for each pixel in the red-eye candidate region, the color of the pixel is located inside the ellipsoid, Red-eye pixel specifying means for specifying the pixel as a red-eye pixel.

  Thereby, the image processing apparatus can detect the red-eye pixel in which the red-eye phenomenon occurs with high accuracy.

  According to a preferable aspect in the first aspect, the ellipsoid defining means further includes the color of the definition pixel group in which the redness value corresponds to the upper predetermined ratio among the pixel group in the red-eye candidate region. The ellipsoid is defined based on the values of the elements.

  According to this aspect, the image processing apparatus can detect the red-eye pixel in which the red-eye phenomenon has occurred with higher accuracy.

  According to a preferred embodiment in the first aspect, the first and second elements of the elements are two redness values having different calculation methods based on the color of the pixel, The element is one of the brightness, luminance, and red value of the color of the pixel.

  According to this aspect, the image processing apparatus can concentrate the distribution of red-eye pixels in which the red-eye phenomenon occurs in the three-axis space, and can detect the red-eye pixels with high accuracy.

  According to a preferable aspect in the first aspect, the ellipsoid defining means further defines a center point of the ellipsoid based on an average value of values of the elements of the definition pixel group, and the definition pixel Three diameters of the ellipsoid are defined based on the standard deviation value of the values of each element of the group.

  According to this aspect, the image processing apparatus can define an ellipsoid that separates red-eye pixels and non-red-eye pixels with high accuracy in a three-axis space.

  According to a preferable aspect in the first aspect, the ellipsoid defining means further rotates an axis whose center is the center point of the ellipsoid and the third element is the lightness or red value. A rotation angle of the ellipsoid serving as an axis is defined based on a ratio of average values of the first and second redness values of the definition pixel group.

  According to this aspect, the image processing apparatus can define an ellipsoid that separates red-eye pixels and non-red-eye pixels with higher accuracy in the three-axis space.

  According to a preferred embodiment in the first aspect, the ellipsoid defining means further has a first ellipsoid parameter defining the default ellipsoid, and the first ellipsoid parameter is defined as the definition. The second ellipsoid parameter is obtained by changing the value based on the average value and standard deviation value of each element of the pixel group.

  According to this aspect, the image processing apparatus only needs to change the parameters that need to be changed, and thus can define an ellipsoid that efficiently separates red-eye and non-red-eye pixels.

  According to a preferable aspect in the first aspect, the ellipsoid defining means further changes each rotation angle of the ellipsoid from 0 ° with each axis of the first and second elements as a rotation axis. do not do.

  According to this aspect, by setting the rotation angle parameter corresponding to the two axes to 0 °, the image processing apparatus can omit the parameter to be changed and can suppress the processing load.

  According to a preferable aspect in the first aspect, the image processing apparatus further includes a correction unit that performs correction for suppressing saturation of the pixel for the pixel specified by the red-eye pixel specifying unit. The red-eye pixel specifying unit further generates correction coefficient information indicating a correction degree for each pixel, and the correction coefficient of the pixel located inside the ellipsoid is a first correction indicating the maximum correction degree. The correction coefficient of the pixel located outside the ellipsoid is a second correction coefficient that is smaller than the first correction coefficient and includes 0, and the correction means is based on the correction coefficient of the pixel. Correction for suppressing the saturation is performed.

  According to this aspect, the image processing apparatus can specify and correct a pixel located inside the ellipsoid as a red-eye pixel.

  According to a preferred aspect of the first aspect, the red-eye pixel specifying unit further corrects the pixel located within the ellipsoid and having a distance within a predetermined range from a boundary of the ellipsoid. A coefficient is set as a third correction coefficient that is gradually decreased from the first correction coefficient to the second correction coefficient according to the distance, and the correction coefficient of the pixel located outside the ellipsoid is determined according to the distance. The fourth correction coefficient is gradually increased from the second correction coefficient to the first correction coefficient.

  According to this aspect, the image processing apparatus is a pixel whose distance from the boundary of the ellipsoid is within a predetermined range, the pixel positioned outside the ellipsoid while suppressing the correction degree of the pixel positioned inside the ellipsoid. Although the degree of correction is small, it can be the target of correction.

In order to achieve the above object, according to the second aspect of the present invention, a computer that causes a computer to execute image processing for identifying a red-eye pixel from a red-eye candidate region including a red-eye in which a red-eye phenomenon occurs in a color image. In a readable image processing program, for a color image that identifies a red-eye pixel , a pixel group in the red-eye candidate region in a three-axis space in which at least one color element is a redness value with a color element as an axis An ellipsoid defining step for defining an ellipsoid separating the red-eye pixel and the non-red-eye pixel based on the value of the color element; and for each pixel in the red-eye candidate region, the color of the pixel is the ellipse When located inside the body, a red-eye pixel specifying step of specifying the pixel as a red-eye pixel is executed.

1 is a perspective view illustrating a configuration of a printer 1 according to an embodiment of the present invention. FIG. 2 is a block diagram illustrating a functional configuration example of the printer 1 in FIG. 1. It is the figure which showed typically the printing process target image TI. It is a figure showing red-eye candidate area | region Pe1. It is a figure showing the pixel distribution of the red-eye candidate area | region in the XY-axis plane of triaxial space. It is a figure showing the pixel distribution of the red-eye candidate area | region in the XZ-axis plane of triaxial space. It is a figure showing the pixel distribution of the red-eye candidate area | region in the YZ-axis plane of triaxial space. It is a figure showing the ellipsoid which contains distribution of the red-eye pixel in 3-axis space. It is a flowchart figure which shows the flow of an automatic red eye correction process. It is a figure showing the ellipsoid parameter for defining an ellipsoid. It is a flowchart figure showing the detail of an ellipsoid definition process. It is an example figure showing the straight line L1 and the straight line L2 in triaxial space. It is a figure showing the example of an ellipsoid parameter, and its ellipsoid. It is a flowchart figure which shows the flow of the red-eye pixel specific process Z2. It is a figure showing the relationship between the distance from an ellipsoid, and a correction coefficient. It is a figure showing the correction coefficient in a red-eye candidate area | region. It is a flowchart figure which shows the flow of the correction process Z3.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the technical scope of the present invention is not limited to these embodiments, but extends to the matters described in the claims and equivalents thereof.

  The image processing apparatus according to the present embodiment uses a red-eye based on the element value of a pixel group in a red-eye candidate area having a red-eye pixel in a three-axis space including a color element as an axis and at least a redness value as an element. An ellipsoid that separates pixels and non-red-eye pixels is defined, and a pixel located inside the ellipsoid is specified as a red-eye pixel.

[Description of Printer 1 Configuration]
FIG. 1 is a block diagram illustrating a hardware configuration example of a printer 1 according to an embodiment of the present invention. The printer 1 is a color ink jet printer or color laser printer compatible with so-called direct printing, in which an image is printed based on image data acquired from the memory card 18 or the like. The printer 1 automatically detects a red-eye portion of a color image to be printed, and corrects the portion to a natural color (for example, black) (hereinafter referred to as an automatic red-eye correction function). have.

  The printer 1 includes a CPU 11, an internal memory 12, an operation unit 13, a display unit 14, a printer engine 15, and a card interface (card I / F) 16.

  A CPU (central processing unit) 11 controls each unit and executes various processes according to a program stored in the internal memory 12.

  For example, when the automatic red-eye correction mode is set, the CPU 11 selects an eye such as a person, a dog or other pet from a color image to be printed (hereinafter referred to as a print processing target image TI) prior to printing. A process for detecting an area including an image or an area around a point designated by the user (hereinafter referred to as a detection process Z0) is executed. Further, the CPU 11 executes a process (hereinafter referred to as an ellipsoid definition process Z1) that defines an ellipsoid that includes red-eye pixels in the triaxial space based on the detected area (hereinafter referred to as a red-eye candidate area Pe1). To do. Then, the CPU 11 specifies a pixel located inside the ellipsoid in the three-axis space as a red-eye pixel (hereinafter referred to as a red-eye pixel specifying process Z2), and sets the color of the specified red-eye pixel to natural. A process of replacing with a color (for example, black) (hereinafter referred to as correction process Z3) is executed.

  The internal memory 12 includes a ROM (Read Only Memory) that stores various programs executed by the CPU 11 and various data, and a RAM (Random Access Memory) that temporarily stores programs and data to be executed by the CPU 11. Has been.

  The operation unit 13 includes one or a plurality of buttons and a touch panel, and notifies the CPU 11 of the operation content by the user.

  The display unit 14 includes a liquid crystal display and displays an image corresponding to the display data supplied from the CPU 11.

  The printer engine 15 is a printing mechanism that performs printing based on print data supplied from the CPU 11.

  The card interface (I / F) 16 is an interface for exchanging data with the memory card 18 inserted in the card slot 17. In this example, image data as RGB data is stored in the memory card 18, and the printer 1 acquires the image data stored in the memory card 18 via the card interface 16.

  The components of the printer 1 are connected to each other via a bus 19.

  The printer 1 may further include an interface for performing data communication with other devices (for example, a digital still camera or a computer).

[Description of Functional Configuration Example of Printer 1]
FIG. 2 is a block diagram illustrating a functional configuration example of the printer 1 for executing the detection process Z0, the ellipsoid definition process Z1, the red-eye pixel specifying process Z2, and the correction process Z3. This function can be realized by the CPU 11 executing a predetermined program stored in the internal memory 12.

  The image reading unit 21 reads the print processing target image TI from, for example, the memory card 18 via the card I / F 16.

FIG. 3 is a diagram schematically showing the print processing target image TI. Here, it is assumed that the red eye phenomenon occurring in the left eye portion 31 of the person at the center of the print processing target image TI is the processing target.

  The red-eye candidate area acquisition unit 22 performs a detection process Z0 on a replica of the print processing target image TI read by the image reading unit 21 (hereinafter also simply referred to as a print processing target image TI). Specifically, for example, a known detection method such as a pattern matching method using a template representing a basic shape of a person's eyes (for example, Japanese Patent Application Laid-Open No. 2006-279460) can be used to detect pupils, irises, white eyes, A process of acquiring a region including a part such as the eyes and the corners of the eyes is executed. Further, the red-eye candidate area acquisition unit 22 executes a process of detecting an area including a reddish pupil and iris from the area acquired by a known detection method, and acquires the left-eye part as the red-eye candidate area Pe1.

  FIG. 4 is a diagram illustrating the red-eye candidate region Pe1. The red-eye candidate region Pe1 in the figure is a rectangular region including the pupil and iris in which the red-eye phenomenon occurs in the print processing target image TI in FIG. 3 and the surrounding white eyes, eyes, and eye corners.

  5, 6, and 7 are elements of “R value−B value (X axis)”, “R value−G value (Y axis)”, and “G value (Z axis)” based on the RGB color space. It is an example figure showing distribution of a pixel group in red eye candidate field Pe1 in a plane of 3 axis space used as a value. Both “R value−B value” and “R value−G value” are redness values, and are redness values with different calculation methods. The “G value” is the value of the G component in the RGB color space. In the figure, a black square mark (♦ mark) indicates a pixel in which a red-eye phenomenon occurs, and a white square mark (□ mark) indicates other pixels.

  FIG. 5 is an example showing a distribution of pixels in the red-eye candidate region Pe1 on the “R value−B value (X axis)” and “R value−G value (Y axis)” plane in the triaxial space. In the plane of the figure, non-red-eye pixels (□) are concentrated in the lower left part of the plane. On the other hand, the red-eye pixels (♦ marks) are distributed in the shape of an ellipse 51 centered around the center of the plane. Since the red-eye pixel is a pixel with a high redness value, the two-axis values (redness values) are higher than the non-red-eye pixel on a plane centered on two different redness values. Concentrated and distributed.

  FIG. 6 is an example showing the distribution of the pixel group in the red-eye candidate region Pe1 on the “R value−B value (X axis)” and “G value (Z axis)” plane in the triaxial space. In the plane of the figure, the non-red-eye pixels (□) are portions where the redness value “R value−B value” is low and the “G value” is distributed in the range of 20 to 150. In contrast, the red-eye pixels (♦ marks) have a high redness value “R value−B value” and a “G value” distributed in the range of 0-80. That is, the red-eye pixel has a higher redness value “R value-B value” in the “R value-B value” “G value” plane than the non-red-eye pixel, and has a “G value” representing brightness. The values are distributed in a lower and narrower range 51.

  FIG. 7 is an example showing the distribution of the pixel group in the red-eye candidate region Pe1 on the “R value−G value (Y axis)” and “G value (Z axis)” plane in the triaxial space. In the plane of the figure, the non-red-eye pixels (□) are portions where the redness value “R value−G value” is low, and the “G value” is distributed in the range of 20 to 150. On the other hand, the red-eye pixels (♦ marks) have a high redness value “R value−G value” and are distributed in a range of “G value” from 0 to 80. That is, the red-eye pixel has a higher redness value “R value-G value” in the “R value-G value” “G value” plane than the non-red-eye pixel, and has a “G value” representing brightness. The values are distributed in a lower and narrower range 51.

  5 to 7, the red-eye pixels are distributed in an ellipse 51 in the X-axis-Y-axis plane, and the ellipse is in the X-axis-Z axis in FIG. 6 and the Y-axis-Z axis plane in FIG. 51 red-eye pixels are distributed in a certain value range on the Z-axis. That is, it can be seen that the red-eye pixels in the red-eye candidate region Pe1 are distributed in the range of the shape of the ellipsoid 51 in the three-axis space of FIGS.

  FIG. 8 is a diagram showing an ellipsoid 81 containing the distribution of red-eye pixels in each plane of FIGS. 5 to 7 on a triaxial space. The elements of each axis are the same as in FIGS. As shown in FIG. 8, for example, in a three-axis space having “R value-B value”, “R value-G value”, and “G value” as elements of the axes, the red-eye pixel is a space of an ellipsoid 81 as shown in FIG. It tends to be concentrated in the inside.

  Therefore, the image processing apparatus according to the present embodiment uses the ellipsoid in which the red-eye pixels are distributed in a three-axis space with the color element as an axis as a reference for the pixels in the red-eye candidate region Pe1 as red-eye pixels and non-red-eye pixels. Divided into pixels. Further, the image processing apparatus performs red-eye correction on the red-eye pixel based on the positional relationship with the ellipsoid, for example, the distance between the boundary of the ellipsoid on the three-axis space and the pixel.

  Note that the shape and position of an ellipsoid representing the distribution of red-eye pixels in the three-axis space vary depending on the color characteristics of the red-eye pixel group included in the red-eye candidate region Pe1. Therefore, the image processing apparatus according to the present embodiment uses an ellipsoid that separates red-eye pixels and non-red-eye pixels in a three-axis space in which at least one color element is a redness value with a color element as an axis. This is defined based on the element value of the color of the pixel group in the candidate area Pe1.

  Returning to FIG. 2, the ellipsoid defining unit 23 defines an ellipsoid in a three-axis space with the element as an axis based on the color element of the red-eye candidate area Pe <b> 1 acquired by the red-eye candidate area acquiring unit 22. . The ellipsoidal definition unit 23 acquires the element value of the color of the red-eye candidate area Pe1, and acquires an ellipsoid parameter that defines the ellipsoid based on the acquired element value. Then, the ellipsoid definition unit 23 defines an ellipsoid in the three-axis space based on the calculated ellipsoid parameters. Details of the process of defining an ellipsoid will be described later as ellipsoid definition process Z1.

  Then, the red-eye pixel specifying unit 24 performs a red-eye pixel specifying process Z2 that specifies a red-eye pixel based on the ellipsoid in the three-axis space generated by the ellipsoid defining unit 23. The red-eye pixel specifying unit 24 specifies a pixel located at the coordinates inside the ellipsoid of the triaxial space as a red-eye pixel for each pixel in the red-eye candidate region Pe1. Specifically, the red-eye pixel specifying unit 24 generates correction coefficient information based on the positional relationship between each pixel and the ellipsoid, and stores the correction coefficient information in the correction coefficient map Ma. The correction coefficient is a value representing the degree of red-eye correction.

  The correction unit 25 corrects the color saturation of the pixel in the red-eye candidate region Pe1 based on the correction coefficient information stored in the correction coefficient map Ma generated by the red-eye pixel specifying unit 24 (corrects the red eye). Processing Z3 is executed to generate print data.

  Details of the detection process Z0, ellipsoid definition process Z1, red-eye pixel specifying process Z2, and correction process Z3 will be described below.

  FIG. 9 is a flowchart showing the flow of automatic red-eye correction processing. In step S11, the red-eye candidate area acquisition unit 22 executes the detection process Z0 on the print processing target image TI read by the image reading unit 21, and acquires the red-eye candidate area Pe1. The red-eye candidate region Pe1 is a region including pixels in which the red-eye phenomenon occurs as shown in FIG. However, the red-eye candidate region Pe1 does not necessarily need to be a rectangular region as shown in FIG. 4, and does not need to include the entire eye. The red-eye candidate region Pe1 only needs to include a region where the red-eye phenomenon occurs.

  Subsequently, in step S12, an ellipsoid definition process Z1 is performed.

[Ellipsoid Definition Process Z1: Explanation of Ellipsoid Definition Process]
The ellipsoidal definition unit 23 calculates the redeye pixel and the non-redeye pixel in the three-axis space based on the element value of the color of the redeye candidate region Pe1 acquired by the redeye candidate region acquisition unit 22 as the ellipsoid definition process Z1. An ellipsoid to be separated is defined (S12).

  In addition, the element values of the respective colors in the triaxial space in the present embodiment are two redness values “R value−B value (X axis)” “R” with different calculation methods, as in FIGS. Value-G value (Y axis) "and" G value (Z axis) "indicating the brightness. However, the element value of each color in the triaxial space is not limited to the above example. Details will be described later.

FIG. 10 is a diagram illustrating ellipsoid parameters for defining an ellipsoid. The ellipsoid parameters include an ellipsoid center point (x ₀ , y ₀ , z ₀ ) 101, three diameters of the ellipsoid (X diameter, Y diameter, Z diameter), and rotation angles (θx, θy, θz).

First, the center point (x ₀ , y ₀ , z ₀ ) 101 is the coordinates of the center of the ellipsoid. The X diameter 102 is half the length of the ellipsoid in the X axis direction, the Y diameter 103 is half the length of the ellipsoid in the Y axis direction, and the Z diameter 104 is the ellipsoid in the Z axis direction. Half the length. Θz105 is the rotation angle of the ellipsoid having the center point (x ₀ , y ₀ , z ₀ ) 101 of the ellipsoid as the rotation center and the Z axis as the rotation axis. Similarly, θx is the rotation angle of the ellipsoid with the center point (x ₀ , y ₀ , z ₀ ) of the ellipsoid as the rotation center and the X axis as the rotation axis, and θy is the ellipsoid with the Y axis as the rotation axis. Rotation angle.

  FIG. 11 is a flowchart showing details of the ellipsoid definition process. The ellipsoid definition unit 23 acquires the ellipsoid parameters shown in FIG. 10 and generates an ellipsoid based on the parameters.

  First, the ellipsoidal definition unit 23 detects a definition pixel group whose redness value corresponds to the upper predetermined ratio among the pixel groups in the red-eye candidate region Pe1 (S21). Specifically, the ellipsoidal definition unit 23 acquires the redness values “R value−B value” and “R value−G value” of the pixels in the red eye candidate region Pe1, and the redness value “R value−B value”. ”Corresponds to the upper 10% of the red eye candidate area Pe1 (hereinafter, the definition pixel group RB), and the redness value“ R value−G value ”corresponds to the upper predetermined ratio of the red eye candidate area Pe1. Each pixel group to be detected (hereinafter, defined pixel group RG) is detected. In other words, the ellipsoidal definition unit 23 selects pixel groups (definition pixel groups RB and RG) that have a high redness value and are estimated to have a redeye phenomenon among the pixels in the redeye candidate region Pe1. Detect based on each redness value.

  In the ellipsoidal definition unit 23 in the present embodiment, the ratio of the definition pixel group in the red-eye candidate area Pe1 is about 10%, but the predetermined ratio is not limited to the above 10%. The predetermined ratio depends on, for example, the area ratio between the red-eye candidate area Pe1 and its pupils and irises, that is, the area ratio between the red-eye candidate area Pe1 and the area where the red-eye phenomenon is estimated to occur (red-eye estimation area). It is desirable to be adjusted. For example, when the area of the red-eye estimation region in the red-eye candidate region Pe1 is about 1/20 of the area of the red-eye candidate region, the ellipsoid definition unit 23 adjusts the predetermined ratio to about 5%.

  In addition, the ellipsoidal definition unit 23 acquires the standard deviation value and the average value of the redness value in the red-eye candidate region Pe1, and, for example, sets the redness value “average value + standard deviation value × 2” or more to the red value. You may detect the pixel which has as a degree value as a definition pixel group.

Subsequently, the ellipsoid definition unit 23 acquires the center point parameters (x ₀ , y ₀ , z ₀ ) of the ellipsoid based on the average value of the color element values of the definition pixel groups RB and RG. (S22). Specifically, the ellipsoidal definition unit 23 sets the average value of “R value−B value (X axis)” of the definition pixel group RB as the center point “x ₀ ” on the X axis, and defines the definition pixel group R. An average value of “R value−G value (Y axis)” of −G is set as a center point “y ₀ ” on the Y axis. In addition, the ellipsoidal definition unit 23 displays “G” of the definition pixel groups RB and RG.
The average value of “value (Z axis)” is defined as a center point “z ₀ ” on the Z axis.

In this way, the ellipsoidal definition unit 23 calculates the average value of the color element values of the definition pixel groups RB and RG that are estimated to be red-eye pixels as the center point (x ₀ , y of the ellipsoid). The value of the corresponding element of ₀ , z ₀ ).

  Next, the ellipsoid definition unit 23 determines the ellipsoid diameter parameters (X diameter, Y diameter, Z) based on the dispersion degree (standard deviation value) of the element values of the colors of the definition pixel groups RB and RG. (Diameter) is acquired (S23). First, FIG. 12 will be described.

FIG. 12 is an example showing a straight line L1 and a straight line L2 in the triaxial space. In the figure, a temporary ellipsoid 70 having the center point (x ₀ , y ₀ , z ₀ ) acquired in step 22 is represented by a dotted line. The straight line L1 in the figure is a line connecting the center point (x ₀ , y ₀ , z ₀ ) 101 of the ellipsoid 70 and the origin 100 of the triaxial space, and the straight line L2 is the center point (x ₀ , y of the ellipsoid). ₀ , z ₀ ) 101 and a line perpendicular to L1. The pixel a is an example of the definition coordinate group RB, and the pixel b is an example of the definition coordinate group RG. L1a represents the distance between the pixel a and the straight line L1, and L2a represents the distance between the pixel a and the straight line L2. Similarly, L1b represents the distance between the pixel b and the straight line L1, and L2b represents the distance between the pixel b and the straight line L2.

  Specifically, the ellipsoidal definition unit 23, for example, a distance value (for example, L1a) between the straight line L1 in FIG. 12 and the pixels (for example, the pixel a pixel b) in the definition pixel groups RB and RG. , L1b) is multiplied by a predetermined coefficient α (1 ≦ α ≦ 6) and the value obtained by multiplying the standard deviation value of L1b) by the X diameter. Similarly, the ellipsoidal definition unit 23 determines the distance values (for example, L2a and L2b) between the straight line L2 in FIG. 12 and the pixels (for example, the pixel a pixel b) in the definition pixel groups RB and RG. ) Is multiplied by a predetermined coefficient β (1 ≦ β ≦ 6) as the Y diameter. The ellipsoid defining unit 23 then adds a predetermined coefficient γ (1 ≦ γ ≦ 6) to the standard deviation value of the “G value” of the pixels in the definition pixel groups RB and RG (for example, the pixel a pixel b). The value obtained by multiplying is taken as the Z diameter. It is assumed that the value of each predetermined coefficient has been adjusted in advance.

In this manner, the ellipsoidal definition unit 23, the center point of the pixel and ellipsoid definition pixel group is assumed to be red pixel _{(x 0, y 0, z} 0) the straight line in the street perpendicular relationship Each diameter of the ellipsoid is acquired based on the standard deviation value (dispersion degree) of the distance from L1 and L2 or the standard deviation value (dispersion degree) of the element value of the color of the definition pixel group. Therefore, the larger the degree of dispersion of the distances between the straight lines L1 and L2 and the pixels in the definition pixel group (for example, L1a, L1b, L2a, and L2b), the larger ellipsoid is defined. An ellipsoid is defined. The same applies to the degree of dispersion of the “G value” of the definition pixel group.

  Subsequently, the ellipsoidal definition unit 23 acquires the rotation angle parameters (θx, θy, θz) of the ellipsoid (S24). The ellipsoidal definition unit 23 determines the Z axis rotation angle θz of the ellipsoid, for example, an average value of “R value−B value” of the definition pixel group RB and “R value−” of the definition pixel group RG. Calculation is based on the ratio of the “G value” to the average value. Specifically, the ellipsoidal definition unit 23 in the present embodiment is Arctan (an average value of “R value−G value” of the definition pixel group RG / “R value−B of the definition pixel group RB”). Θz is calculated from the average value of “value”. In the R value-B value (X axis) R value-G value (Y axis) plane, the ellipsoid is rotated in the Z-axis direction based on the calculated θz, so that the distribution of the definition pixel group is made more accurate. An inclusive ellipsoid is defined.

  Then, the ellipsoid definition unit 23 obtains the X-axis rotation angle θx and the Y-axis rotation angle θy as follows, for example. The ellipsoidal definition unit 23 calculates the relative angle between the approximate straight line of the pixels of the definition pixel group RG and the Y axis in the “R value−G value (Y axis)” and “G value (Z axis)” planes of FIG. 7. Is obtained as θx. Similarly, the relative angle between the approximate straight line of the pixels of the definition pixel group RB and the X axis in the “R value−B value (X axis)” and “G value (Z axis)” planes of FIG. get.

  However, in the present embodiment, the approximate straight line and the Y axis of the definition pixel group RG in the “R value−G value (Y axis)” and “G value (Z axis)” planes of FIG. The approximate straight line of the definition pixel group RB and the X axis in the “R value−B value (X axis)” and “G value (Z axis)” planes of FIG. Accordingly, the rotation angle θx and the rotation angle θy in this embodiment are 0 °.

  In this way, the ellipsoidal definition unit 23 determines the ratio of the average values of the element values of the definition pixel groups RB and RG estimated to be red-eye pixels, and the definition pixel group R in each plane. The rotation angle parameter of the ellipsoid is acquired based on the relative angle between the approximate line of the pixels in -B and RG and the corresponding axis.

  As described above, θx and θy in this embodiment are 0 °. As described above, since there is an axis that does not rotate among the three axes, the ellipsoidal definition unit 23 can reduce the load of the calculation process in the red-eye pixel specifying process Z2 described later. Thereby, the ellipsoidal definition part 23 can specify a red-eye pixel more efficiently.

  In addition, the ellipsoidal definition unit 23 uses two elements among elements corresponding to the respective axes of the triaxial space as color elements such as a redness value and a red value, as in the present embodiment. (For example, X-axis, Y-axis) By using one element as an element representing brightness or luminance (for example, Z-axis), the image processing apparatus has two axes of color elements (X-axis, Y-axis). Can be set to 0 °. Thereby, the acquisition processing of the rotation angles θx and θy can be omitted, and the load of calculation processing in the above-described red-eye pixel specifying processing Z2 can be suppressed.

FIG. 13 is a diagram illustrating an example of an ellipsoid parameter and the ellipsoid. The ellipsoid definition unit 23 defines an ellipsoid as shown in FIG. 12 based on the acquired ellipsoid parameters. In the ellipsoid of FIG. 12, the center point 101 is (x ₀ , y ₀ , z ₀ ) = (110, 120, ₀ ), the X diameter 102 is “100”, the Y diameter 103 is “30”, and the Z diameter 104 Is an ellipsoid with "50". Further, the rotation angles θx and θy of the ellipsoid are “0 °”, and θz105 is “45 °”. In addition, although the length of each diameter differs in the ellipsoid of the figure, the ellipsoid may be a spheroid having two equal diameters among three diameters or a sphere having all equal diameters.

  As described above, the ellipsoidal definition unit 23 according to the present embodiment determines the ellipsoidal parameter based on the element value of the color of the definition pixel group that is estimated to be the redeye pixel in the redeye candidate area Pe1. Acquire and define an ellipsoid in the 3-axis space. As a result, the ellipsoidal definition unit 23 separates the red-eye pixels and the non-red-eye pixels in the red-eye candidate region Pe1 based on the color characteristics of the definition pixel group in the red-eye candidate region Pe1. A body can be defined.

  In addition, in step S21, the ellipsoid defining unit 23 in the present embodiment has the redness values “R value−B value” and “R value−G value” in the upper 10% of the red eye candidate region Pe1 in the triaxial space. Are detected as a definition pixel group. However, the ellipsoidal definition unit 23 has a redness value different from the redness value (“R value−B value” “R value−G value”) of the element in the triaxial space (for example, “R × R− ( B + G) ”) may be detected as a definition pixel group (hereinafter referred to as a definition pixel group R × R− (B + G)), which is the upper predetermined ratio of the red-eye candidate region Pe1.

In that case, the ellipsoidal definition unit 23 averages the element values (“R value−B value”, “R value−G value”, “G value”) corresponding to each axis of the definition pixel group R × R− (B + G). The value is the center point (x ₀ , y ₀ , z ₀ ) of the ellipsoid (S22), the standard deviation value of the distance between the pixels in the definition pixel group R × R− (B + G) and the straight lines L1, L2, and “ Each diameter of the ellipsoid is acquired based on the standard deviation value of “G value” (S23). Then, the ellipsoidal definition unit 23 determines the ratio of the average values of “R value−B value” and “R value−G value” of the definition pixel group R × R− (B + G), or the definition pixel group R × in each plane. A rotation angle is acquired based on the relative angle of the axis corresponding to the approximate straight line of R− (B + G) (S24).

  The ellipsoid defining unit 23 prepares ellipsoid parameters for defining a default ellipsoid based on a large number of red eye candidate areas Pe1, and sets only the parameters that need to be changed to the color of each red eye candidate area Pe1. You may acquire based on an element value. Thereby, the ellipsoidal definition part 23 can omit the acquisition process for parameters that do not need to be changed, and can more efficiently identify the red-eye pixel.

The value of the center point 101 (x ₀ , y ₀ , z ₀ ) = (110, 120, ₀ ) of the ellipsoid in FIG. 12 is “0”. It is assumed that the value of the axis is also “0”. In this way, by setting the default ellipsoid parameters and the elements of the axes so that some of the ellipsoid parameters do not need to be changed, the ellipsoid definition unit 23 performs part of the ellipsoid parameter change processing. It can be omitted and an ellipsoid can be defined efficiently.

[Red-eye pixel specifying process Z2: Description of red-eye pixel specifying process]
After defining the ellipsoid (S12 in FIG. 9), the red-eye pixel specifying unit 24 subsequently executes a red-eye pixel specifying process Z2 in step S13 in FIG.

  The red-eye pixel specifying unit 24 in the present embodiment generates a correction coefficient map Ma as information indicating a red-eye pixel specifying result. The correction coefficient map Ma has a storage unit corresponding to each pixel of the red-eye candidate region Pe1. Assuming that the storage unit is a pixel, the correction coefficient map Ma has at least the same pixel size as the red-eye candidate region Pe1. In the present embodiment, the red-eye pixel specifying unit 24 sets “correction coefficient> 0” for a target pixel for red-eye correction, “correction coefficient = 0” for a pixel that is not a target for red-eye correction, and a corresponding correction coefficient map. The data is stored in each storage unit (that is, each pixel) of Ma.

  FIG. 14 is a flowchart showing the flow of the red-eye pixel specifying process Z2. First, the red-eye pixel specifying unit 24 selects one pixel of interest (hereinafter referred to as a pixel of interest) from the red-eye candidate region Pe1 (S31), acquires a correction coefficient of the pixel of interest, and stores it in the correction coefficient map Ma. (S32). Specifically, the red-eye pixel specifying unit 24 in the present embodiment substitutes the coordinates (x, y, z) of the target pixel in the following formula (1), and the pixel is positioned inside the ellipsoid. It is determined whether or not to do.

In the above formula (1), (x ₀ , y ₀ , z ₀ ) is the center point of the ellipsoid, and a is the X of the ellipsoid.
The diameter, b is the Y diameter of the ellipsoid, c is the Z diameter of the ellipsoid, and θx, θy, and θz represent the rotation angles corresponding to the respective axes. If the coordinates (x, y, z) of the target pixel are substituted into the above equation (1), the distance “E value” from the center point of the ellipsoid is obtained. When the E value of the target pixel is “E value = 1”, it means that the target pixel is located on the boundary of the ellipsoid in the three-axis space. Further, when the E value of the target pixel is “E value <1”, it indicates that it is located inside the ellipsoid, and when “E value> 1”, it indicates that it is located outside the ellipsoid.

  Formula (1) has a large calculation amount and a high processing load. Therefore, the calculation processing load can be suppressed by having a rotation angle of 0 ° among the three rotation angles. Therefore, when θx and θy are 0 ° in the present embodiment, the red-eye pixel specifying unit 24 can specify the red-eye pixel more efficiently while suppressing the processing load. Subsequently, the red-eye pixel specifying unit 24 acquires a correction coefficient based on the E value.

  FIG. 15 is a diagram illustrating the relationship between the output value “E value” of Expression (1) and the correction coefficient. In the figure, the horizontal axis represents the input value “E value” and the vertical axis represents the correction coefficient. Based on the broken line 151, the red-eye pixel specifying unit 24 assigns a correction coefficient “1” for pixels with “E value ≦ 1” and a correction coefficient “0” for pixels with “E value> 1”.

  By the way, the red-eye pixel specifying unit 24 may adjust the correction coefficient of a pixel located within a predetermined distance range from the boundary surface of the ellipsoid, that is, a pixel located near the boundary of the ellipsoid according to the distance. . For example, even if the pixel is located inside the ellipsoid (“E value ≦ 1”), if the pixel is located within a predetermined distance range from the boundary surface of the ellipsoid, the pixel may not be a red-eye pixel. Even if the pixel is located outside the ellipsoid (“E value> 1”), if the pixel is located within a predetermined distance range from the boundary surface of the ellipsoid, the pixel may be a red-eye pixel. Therefore, the red-eye pixel specifying unit 24 adjusts a correction coefficient of a pixel whose distance from the boundary surface of the ellipsoid is within a predetermined range according to the distance.

  Specifically, the red-eye pixel specifying unit 24 acquires the correction coefficient “S value” by substituting the “E value” calculated by Expression (1) into, for example, Expression (2) below. In the following formula (2), “E” is a value obtained by the formula (1), and “e” indicates a base value of the natural logarithm. “A (gain)” will be described later. Note that the method of changing the correction coefficient of the pixel located near the ellipsoid is not limited to the method of Expression (2).

  A curve 152 in FIG. 15 represents a correspondence relationship between the input value “E value” and the correction coefficient “S value” that is an output value based on Expression (2). In the figure, the horizontal axis represents the input value “E value” and the vertical axis represents the correction coefficient “S value” which is the output value. “A (gain)” in Expression (2) represents the degree of inclination of the curve 151 in the case of “E value = 1” (pixels are located on the boundary surface of the ellipsoid) in FIG. The red-eye pixel specifying unit 24 in the present embodiment sets the value of a to “20”.

  According to Equation (2), for a pixel (about 0.8 ≦ E value <1) that is located inside the ellipsoid and whose distance from the boundary surface of the ellipsoid is within a predetermined range, the correction coefficient “0.5 <S value <1 "is assigned. On the other hand, the correction coefficient “0 <S value <0.5” is set for a pixel (1 <E value ≦ about 1.2) that is located outside the ellipsoid and whose distance from the boundary surface of the ellipsoid is within a predetermined range. Is assigned.

  In this way, the red-eye pixel specifying unit 24 starts from “1” as the distance of the pixel located within the ellipsoid within the predetermined range and the distance from the boundary surface of the ellipsoid decreases. Assign a decreasing correction factor. On the other hand, the red-eye pixel specifying unit 24 increases the pixel from the boundary of the ellipsoid from “0” as the distance becomes shorter with respect to a pixel located outside the ellipsoid within a predetermined range. Assign a correction factor.

  As a result, pixels that may be red-eye pixels even if they are located outside the ellipsoid are subject to correction, although the degree of correction is small, and may be non-red-eye pixels even if they are located inside the ellipsoid A certain pixel has a small degree of correction.

  FIG. 16 is an example showing a correction coefficient in the red-eye candidate region Pe1. 161 in the figure is the correction coefficient when the correction coefficient of the pixel located near the boundary surface of the ellipsoid is not adjusted (151 in FIG. 15), and 162 is the adjustment of the correction coefficient of the pixel located near the boundary surface of the ellipsoid. It is an example figure showing the correction coefficient in the case of doing (152 of Drawing 15). In the drawing, a black portion represents a pixel having “correction coefficient = 1”, and a white portion represents a pixel having “correction coefficient = 0”. The shaded portion represents a pixel where “0 <correction coefficient <1”.

  In FIG. 16, the correction coefficient gradually changes without being separated into a correction target pixel and a non-correction target pixel. As a result, the image processing apparatus according to the present embodiment allows the red-eye candidate region Pe1 including many pixels located near the boundary surface of the ellipsoid to be corrected and non-corrected by the red-eye correction process. It is possible to prevent the image from being converted into an unnatural image in which the difference is noticeable. Similarly, a certain amount of red-eye correction is performed on a pixel that is not affected by the red-eye phenomenon even if it is not the pupil or iris in the red-eye candidate region Pe1.

  Incidentally, the red-eye pixel specifying unit 24 may change the distance range from the boundary surface of the ellipsoid whose adjustment coefficient is to be adjusted based on, for example, the length of the diameter of the ellipsoid. For example, when the diameter of the ellipsoid is shorter than a predetermined value, the red-eye pixel specifying unit 24 considers that the ellipsoid is small and the distribution of red-eye pixels is concentrated, and conversely reduces the distance range. When the value is longer than the value, the ellipsoid is considered to be large and the degree of dispersion of the distribution of red-eye pixels is regarded as large, and the distance range is expanded.

  Specifically, when reducing the distance range, the red-eye pixel specifying unit 24 changes the value of “a” in Expression (2) from “20” to “30”, for example, by changing the curve (152 in FIG. 15). ) Make the slope of steep. On the other hand, when expanding the distance range, the red-eye pixel specifying unit 24 changes the slope of the curve (152 in FIG. 15) by changing the value of “a” in Expression (2) from “20” to “10”, for example. To loosen. A dotted curve 153 in FIG. 15 represents a curve when the value of “a” is “10”. The curve 153 in the figure has a gentler slope than the curve 152 (a = 20), and has a wide E value range where the correction coefficient is “0 <S value <1”.

  When the red-eye pixel specifying unit 24 acquires the correction coefficient, the red-eye pixel specifying unit 24 stores the correction coefficient in the storage unit of the correction coefficient map Ma corresponding to the target pixel. Subsequently, the red-eye pixel specifying unit 24 determines whether or not all the pixels in the red-eye candidate region Pe1 have been selected as the target pixel (S33). If there is a pixel that has not yet been selected as the target pixel (NO in S33), the red-eye pixel specifying unit 24 returns to step S31, selects another pixel as the target pixel, and similarly executes the processing from step S32 onward. To do.

  When all the pixels in the red-eye candidate region Pe1 are selected as the target pixel (YES in S33), the correction coefficient map Ma is generated. In the correction coefficient map Ma in the present embodiment, a correction coefficient “0 <S value ≦ 1” is stored for the correction target pixel, and a correction coefficient “0” is stored for the non-correction target pixel.

  As described above, the red-eye pixel specifying unit 24 generates the correction coefficient map Ma corresponding to each pixel of the red-eye candidate region Pe1 based on the separation curve (S13 in FIG. 9).

[Description of Correction Process Z3]
Subsequently, the correction unit 25 executes the correction process Z3 based on the correction coefficient map Ma in step S14 of FIG.

  FIG. 17 is a flowchart showing the flow of the correction process Z3. First, the correction unit 25 selects a pixel of interest from the red-eye candidate region Pe1 (S41), and reads a correction coefficient stored in the storage unit corresponding to the pixel of interest from the correction coefficient map Ma (S42).

  Subsequently, the correction unit 25 performs a red-eye correction process for suppressing the saturation of the color of the target pixel according to the read correction coefficient (S43). Specifically, when the correction coefficient of the pixel of interest is “0.5”, for example, the correction unit 25 has a degree of half of the amount of change in saturation when the correction coefficient is “1”. Make corrections to reduce saturation.

  Then, the correction unit 25 determines whether or not all the pixels in the red-eye candidate region Pe1 have been selected as the target pixel (S44). If there is still a pixel that has not been selected as the target pixel (NO in S44), the process returns to step S41, and the correction unit 25 selects another pixel as the target pixel, and similarly executes the processing from step S42 onward. . On the other hand, when all the pixels in the red-eye candidate area Pe1 are selected (YES in S44), the correction unit 25 ends the correction process.

  As described above, the image processing apparatus according to the present exemplary embodiment uses the ellipsoid in the three-axis space that separates the red-eye pixel and the non-red-eye pixel based on the color element value of the pixel group in the red-eye candidate region. Define and identify each pixel in the red-eye candidate region as a red-eye pixel when the pixel is located inside the ellipsoid. The image processing apparatus according to the present embodiment performs the automatic red-eye correction process (S11 to S14 in FIG. 9) on all red-eye candidate areas Pe1 in the print processing target image TI.

  In this way, by defining an ellipsoid that separates the distribution of red-eye pixels and non-red-eye pixels in the three-axis space, the image processing apparatus determines red-eye pixels and non-red-eye pixels based on an ellipse on a biaxial plane. Compared to the case of separation, the distribution of red-eye pixels can be defined in a multifaceted manner. Therefore, the image processing apparatus can specify the red-eye pixel with higher accuracy.

  In addition, since the image processing apparatus in the present embodiment defines an ellipsoid based on the color element values in the red-eye candidate area Pe1, the position of the ellipsoid based on the color features in the red-eye candidate area Pe1 and The shape can be defined, and the red-eye pixel can be specified with high accuracy.

  Furthermore, the image processing apparatus according to the present embodiment defines an ellipsoid based on the element value of the color of the definition pixel group which is a pixel in the red-eye candidate region Pe1 and whose redness value is a higher ratio. As a result, the image processing apparatus defines the ellipsoid based on the color characteristics of the pixel group estimated to be the red-eye pixel in the red-eye candidate region Pe1, and thus can identify the red-eye pixel with higher accuracy.

  In the image processing apparatus according to the present embodiment, the element values in the three-axis space are “R value−B value”, “R value−G value”, and “G value”, but the present invention is not limited to this example. is not. For example, the image processing apparatus may assign the following values to the element values in the three-axis space.

For example, the image processing apparatus changes “G value” to another element value among the three-axis elements “R value−B value”, “R value−G value”, and “G value” in the present embodiment. May be. For example, the image processing apparatus uses “Y value” representing the luminance in the YCbCr color space or “L value” representing the lightness in the L * a * b color space instead of the “G value” representing the lightness.

  In this case, as in the present embodiment, the image processing apparatus uses two redness values (“R value-B value”, “R value— G-value ") and red-eye pixels having a certain range of brightness or luminance are separated by" Y value "or" L value ". Accordingly, the image processing apparatus can define an ellipsoid that separates red-eye pixels even in the above-described three-axis space.

  The image processing apparatus may use an “R value” (red value) in the RGB color space instead of the “G value”. In this case, the image processing apparatus separates the red-eye pixel having a characteristic with a high redness value by two types of redness values (“R value−B value” and “R value−G value”) having different calculation methods, and further The red-eye pixels having a certain range of red values are separated by “R value”. Accordingly, the image processing apparatus can similarly define an ellipsoid that separates red-eye pixels in a three-axis space that uses an “R value” representing a red value instead of a “G value”.

  Further, the image processing apparatus, for example, “R value / G value”, “R value / B value”, “G value / R value” instead of the redness values “R value−B value” and “R value−G value”. Another redness value such as “value” “B value / R value” may be used. However, when “R value / G value” or “R value / B value” is used as the redness value, the red eye pixel has a larger R value than the G value and the B value, so the range of the redness value is widened. It will pass and the ellipsoid will become larger. In this case, the image processing apparatus uses the “G value / R value” and the “B value / R value” as the redness value, so that the range of the redness value of the redeye pixel falls within a certain range, and the redeye pixel is included. An ellipsoid can be defined.

  In addition, the image processing apparatus has been described using an example of using two types of redness values, brightness, luminance, and red value as elements of the three-axis space, but the redness value, red value, and brightness ( (Luminance) may be a triaxial element. In this case, the image processing apparatus separates red-eye pixels having a certain range of red values and high redness values by the redness value and the red value, and further selects red-eye pixels having a certain range of brightness or luminance from the brightness or luminance. Separate by. Accordingly, the image processing apparatus can similarly define an ellipsoid that separates red-eye pixels in the above-described three-axis space.

  The image processing apparatus may use “R value × R value / G value × B value”, “2 × R value−G value−B value”, or the like in which the redness is more emphasized as the redness value. . The image processing apparatus can more clearly separate the red-eye pixel and the non-red-eye pixel in the three-axis space by using the element value that enhances the redness.

  Further, the image processing apparatus converts the three element values in the triaxial space into “Y value, Cb value, Cr value” in the color space YCbCr, or “L value, a value, b in the color space L * a * b”. Value ”or“ L value, u value, v value ”of the color space L * u * v. In this case, the image processing apparatus performs “Cb value, Cr value” in the color space YCbCr, “a value, b value” in the color space L * a * b, and “u value, v value in the color space L * u * v. ”Are separated from each other based on the element value representing the color tone of the red-eye pixel, and the red-eye pixel having a certain range of lightness or luminance is converted into the“ Y value ”of the color space YCbCr, the color space L * Separation is based on the “L value” of a * b and L * u * v. Thereby, the image processing apparatus can define an ellipsoid that separates red-eye pixels and non-red-eye pixels in the same manner in the three-axis space.

As described above, the image processing apparatus according to the present embodiment can define an ellipsoid that separates red-eye pixels and non-red-eye pixels in a triaxial space having at least a redness value as one of the elements. Further, the image processing apparatus defines a red-eye pixel by defining an ellipsoid in a three-axis space having two types of redness values having different calculation methods and one of brightness, luminance, and red value as axes. The ellipsoid that separates the red-eye pixels and the non-red-eye pixels can be defined with higher accuracy.

  Note that the red-eye phenomenon includes a phenomenon in which the pupil and iris become gold (so-called gold eye) other than when the pupil and iris become red. The image processing apparatus uses the “R value−B value”, “R value / B value”, “B value / R value”, and the like as the element values of the redness value in the three-axis space, thereby causing the gold phenomenon. An ellipsoid representing the distribution of red-eye pixels can be defined to identify the red-eye pixels of the gold. For example, the image processing apparatus according to the present embodiment uses “R value-G value” and “R value-B value” as elements of the three-axis space, so that the red-eye phenomenon occurs, It is also possible to identify a pixel in which the gold phenomenon has occurred.

  Further, the image processing apparatus according to the present embodiment, for example, assigns two types of redness values of the element values in the three-axis space to the X-axis and Y-axis element values, and averages the two types of redness values. By calculating the rotation angle of the Z axis based on the ratio and rotating the ellipsoid, it is possible to define an ellipsoid that is more faithful to the distribution of red-eye pixels.

  Further, the image processing apparatus according to the present embodiment has a certain degree of correction although the degree of correction is small even if the pixel is located outside the ellipsoid and the distance from the boundary of the ellipsoid is within a predetermined range. The correction coefficient is given. As a result, the image processing apparatus can slightly correct pixels that may be involved in the red-eye phenomenon, and can correct red eyes more appropriately.

  Further, the image processing apparatus according to the present embodiment suppresses the correction coefficient even when the pixel is located inside the ellipsoid when the distance from the boundary of the ellipsoid is within a predetermined range. Thereby, the image processing apparatus can suppress the degree of correction of pixels that may be non-red-eye pixels, and can correct red eyes more appropriately.

  Note that the image processing apparatus according to the present embodiment defines an ellipsoid that contains red-eye pixels in a three-axis space. However, the image processing apparatus may define an ellipsoid including non-red-eye pixels on the three-axis space. In this case, for example, the image processing apparatus similarly expresses the distribution of non-red-eye pixels in the three-axis space, using two types of redness values, brightness, luminance, and red value as three-axis elements. Define an ellipsoid. Then, the image processing apparatus specifies a pixel located outside the ellipsoid in the triaxial space as a red-eye pixel.

[Modification]
In the above description, the case where the printer 1 is used has been described as an example. However, the present invention can be applied to other image processing apparatuses such as a computer, a digital video camera, or a digital still camera. That is, it can be implemented in red-eye correction software or the like installed in the image processing apparatus. In this case, a function for generating the correction coefficient map Ma is implemented in the image processing apparatus, the correction coefficient map Ma is provided from the image processing apparatus to the printer or display side, and color correction is executed on the printer or display side. You can also.

  Further, when the functions described above are realized by a computer or the like, a program describing the processing contents of the functions described above is provided. The processing function is realized on the computer by executing the program on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium. Examples of the computer-readable recording medium include a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory. Examples of the magnetic recording device include a hard disk device (HDD), a flexible disk (FD), and a magnetic tape. Examples of the optical disc include a DVD (Digital Versatile Disk), a DVD-RAM, a CD-ROM (Compact Disk ROM), and a CD-R (Recordable) / RW (ReWritable). Magneto-optical recording media include MO (Magneto-Optical disk).

  When distributing the program, for example, portable recording media such as a DVD and a CD-ROM in which the program is recorded are sold. It is also possible to store the program in a storage device of a server computer and transfer the program from the server computer to another computer via a network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can also read a program directly from a portable recording medium and execute processing according to the program. Further, each time the program is transferred from the server computer, the computer can sequentially execute processing according to the received program.

  In the above embodiment, a part of the configuration realized by hardware may be replaced with software, and conversely, a part of the configuration realized by software may be replaced by hardware. .

  The process shown in the flowchart is executed in parallel or individually even if the steps are not necessarily processed in time series, as well as processes performed in time series in the order described. It includes processing that. Further, the processing can be omitted as appropriate.

1 Printer, 11 CPU, 12 Internal memory, 13 Operation unit, 14
Display unit, 15 printer engine, 16 card I / F, 17 card slot, 18 memory card, 21 image reading unit, 22 red eye candidate area acquisition unit, 23 ellipsoidal definition unit, 24 red eye pixel specifying unit, 25 correction unit

Claims (9)

An image processing device for identifying a red-eye pixel from a red-eye candidate region including a red-eye in which a red-eye phenomenon occurs in a color image,
In a three-axis space where a color element is an axis and at least one color element is a redness value for a color image that identifies a red-eye pixel, the value of the color element of the pixel group in the red-eye candidate region An ellipsoid defining means for defining an ellipsoid separating the red-eye pixel and the non-red-eye pixel,
For each pixel in the red-eye candidate region, when the color of the pixel is located inside the ellipsoid, a red-eye pixel specifying unit that specifies the pixel as a red-eye pixel;
Have
The ellipsoid defining means defines the ellipsoid based on a value of the color element of a definition pixel group in which the redness value of the pixel group in the red-eye candidate region corresponds to an upper predetermined ratio. A featured image processing apparatus. In claim 1,
Among the elements, the first and second elements are two kinds of redness values having different calculation methods based on the color of the pixel, and the third element is any one of brightness, luminance, and red value of the color of the pixel. An image processing apparatus characterized by the above. In claim 2,
The ellipsoid defining means defines a center point of the ellipsoid based on an average value of each element of the definition pixel group, and the ellipsoid is defined based on a standard deviation value of each element of the definition pixel group. An image processing apparatus that defines three diameters of a body. In claim 3,
The ellipsoid defining means is further configured to determine a rotation angle of the ellipsoid having a center point of the ellipsoid as a rotation center and an axis whose third element is the lightness or red value as a rotation axis. The image processing apparatus is defined based on a ratio of an average value of the first and second redness values. In any one of Claims 1 thru | or 4,
The ellipsoid defining means includes a first ellipsoid parameter that defines the default ellipsoid, and the first ellipsoid parameter is set to an average value and a standard deviation value of each element of the definition pixel group. An image processing apparatus characterized in that the second ellipsoid parameter is changed based on the second ellipsoid parameter. In claim 4 or 5,
The image processing apparatus according to claim 1, wherein the ellipsoid defining means does not change each rotation angle of the ellipsoid having each axis of the first and second elements as a rotation axis from 0 °. In any one of Claims 1 thru | or 4,
The image processing apparatus further includes:
Correction means for performing correction for suppressing the saturation of the pixel for the pixel specified by the red-eye pixel specifying means;
The red-eye pixel specifying unit further generates correction coefficient information indicating a correction degree for each pixel, and the correction coefficient information of the pixel located inside the ellipsoid is a first correction coefficient indicating a maximum correction degree. And the correction coefficient of the pixel located outside the ellipsoid is a second correction coefficient that is smaller than the first correction coefficient and includes 0,
The image processing apparatus according to claim 1, wherein the correction unit performs correction to suppress the saturation based on the correction coefficient of the pixel. In claim 7,
The red-eye pixel specifying means has a distance within a predetermined range from a boundary of the ellipsoid, and the correction coefficient of the pixel located inside the ellipsoid is determined from the first correction coefficient according to the distance. The third correction coefficient is gradually reduced to the second correction coefficient, and the correction coefficient of the pixel located outside the ellipsoid is changed from the second correction coefficient to the first correction coefficient according to the distance. An image processing apparatus characterized in that the fourth correction coefficient is increased. In a computer-readable image processing program for causing a computer to execute image processing for identifying a red-eye pixel from a red-eye candidate region including a red-eye in which a red-eye phenomenon occurs in a color image,
In a three-axis space where a color element is an axis and at least one color element is a redness value for a color image that identifies a red-eye pixel, the value of the color element of the pixel group in the red-eye candidate region An ellipsoid defining step defining an ellipsoid separating the red-eye and non-red-eye pixels,
For each pixel in the red-eye candidate region, when the color of the pixel is located inside the ellipsoid, a red-eye pixel specifying step of specifying the pixel as a red-eye pixel;
Run
The ellipsoid defining step defines the ellipsoid based on the value of the color element of the definition pixel group in which the redness value corresponds to the upper predetermined ratio among the pixel group in the red-eye candidate region. A featured image treatment program.