JP2002271808A - Program for correcting red-eye in picture, recording medium and red-eye correcting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a red-eye correcting method that can properly correct a red-eye in a picture. SOLUTION: A plurality of feature areas needing red-eye correction are extracted from a picture on the basis of the saturation, lightness and hue being the feature quantities, and a histogram as to each feature quantity of the feature areas is generated (step S32). Parameters such as a minimum value, a maximum value, a mode, a variance, and an average value of the feature quantities are introduced from each histogram and a correction curve with respect to the feature quantities is generated on the basis of the parameters (steps S33, S34). The feature quantities are corrected after that according to the correction curve (step S35). Thus, the feature quantity of each feature area can be properly corrected by using the correction curve for each feature quantity depending on the characteristic of the feature area and red-eye caused in various modes can be properly corrected by taking into account the gradation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a technique for correcting red eyes in an image.

[0002]

2. Description of the Related Art When photographing is performed using a flash, a red-eye phenomenon may occur in which a person's eyes in an image appear shining red or golden. 2. Description of the Related Art Conventionally, a photograph obtained by a silver halide camera is captured by a scanner, or image data obtained by photographing with a digital camera is processed, and red eyes in the image are reduced to a normal eye. correction to the techniques have been proposed.

[0003] As a general technique for correcting red-eye, an area where red-eye exists in an image (hereinafter referred to as a "red-eye area") is replaced with another color of red, or brightness or saturation is changed. The processing of lowering is known. When such simple processing is performed on the red-eye region, the gradation of saturation, lightness, and hue in the eyes is lost, resulting in an unnatural eye. So, for example, JP-2000
In JP-A-76427 and JP-A-2000-134486, the brightness or the saturation is gradually reduced from the periphery to the center of the pupil and the red-eye region.

[0004]

When the lightness or saturation in the red-eye region is corrected in accordance with the position of the pixel to be corrected, the position of the pixel to be corrected and the correction amount need to be appropriately associated. There is. Therefore, for example, when the center of the pupil cannot be detected properly or when the center of the red-eye region is regarded as the center of the pupil even though the center of the red-eye region does not match the center of the pupil, the gradation near the center of the pupil is considered. Becomes unnatural.

[0005] The present invention has been made in view of the above problems, and has as its object to appropriately correct red-eye caused in various modes.

[0006]

Means for Solving the Problems The invention described in claim 1 is a program for correcting the red eye in the image, execution by a computer of the program, the computer, correction target region included in the red eye area Generating a histogram of the feature amount, generating a correction curve based on parameters derived from the histogram, and correcting the feature amount in the correction target area according to the correction curve.

According to a second aspect of the present invention, there is provided the program according to the first aspect, wherein the execution of the program by the computer causes the computer to extract the correction target area from an image based on a feature amount. The process is performed further.

According to a third aspect of the present invention, there is provided the program according to the second aspect, wherein a plurality of correction target areas are extracted from the image, and the histogram is generated for each of the plurality of correction target areas. .

A fourth aspect of the present invention is the program according to any one of the first to third aspects, wherein the parameter is a minimum value, a maximum value, a mode value, an average value, and a variance of the feature amount. Including any of

According to a fifth aspect of the present invention, there is provided the program according to the fourth aspect, wherein the feature amount is a saturation, and the correction curve indicates a maximum value of the saturation as the variance of the saturation is smaller. It has the property of reducing.

[0011] The invention described in claim 6 is the program according to claim 4, it said feature amount is the brightness, the correction curve, reducing an intermediate value of brightness while maintaining maximum brightness It has the property to make it.

According to a seventh aspect of the present invention, there is provided the program according to the fourth aspect, wherein the characteristic amount is saturation or lightness, and a degree of decrease of the intermediate value of the characteristic amount by the correction curve is: It is determined based on the mode of the feature amount.

According to an eighth aspect of the present invention, there is provided the program according to the fourth aspect, wherein the characteristic amount is a hue, and the correction curve converts a mode value of the hue into a predetermined value. It is obtained as a curve.

According to a ninth aspect of the present invention, there is provided the program according to any one of the first to fourth aspects, wherein the histogram is obtained for each of the feature amounts of saturation, lightness, and hue.

[0015] The invention of claim 10 is a program for correcting the red eye in the image, execution by a computer of the program, a plurality extracted to the computer, as red-eye region from the image based on the feature quantity Performing a red-eye correction on the characteristic region, and, before or after the step of performing the red-eye correction, performing a smoothing process on each of the plurality of characteristic regions in accordance with a characteristic region to be processed. Let it run.

[0016] The invention of claim 11 is a program of claim 10, executed by the computer of the program, the computer generates a histogram of feature amount for each of the plurality of characteristic regions And determining the degree of smoothing based on parameters derived from the histogram.

According to a twelfth aspect of the present invention, there is provided a program for correcting red-eye in an image, wherein the computer executes the program by causing the computer to determine a correction target area in the image; a step of comparing the magnitude of a predetermined value of the object area, when the size of the correction target region is smaller than the predetermined value, the step of increasing the resolution of the correction target region with respect to the correction target region And performing the step of performing red-eye correction.

The invention described in claim 13 is a computer-readable recording medium, claims 1 12
The program described in any of the above is recorded.

[0019] The invention according to claim 14, a red eye correction method for correcting red-eye in an image, and generating a histogram of feature amount in the correction target region contained in the red-eye region, derived from the histogram and a step of generating a correction curve based on the parameters, and a step of correcting the characteristic of the correction target region in accordance with the correction curve.

[0020] The invention according to claim 15, a red eye correction method for correcting red-eye in an image, perform red-eye correction for a plurality of characteristic regions extracted as red eye area from the image based on the feature quantity A step of performing, before or after the step of performing the red-eye correction, a step of performing smoothing on each of the plurality of characteristic regions in accordance with a characteristic region to be processed.

According to a sixteenth aspect of the present invention, there is provided a red-eye correction method for correcting red-eye in an image, comprising: determining a correction target area in the image; Performing a step of increasing the resolution of the correction target area when the size of the correction target area is smaller than the predetermined value; and performing a red-eye correction on the correction target area.

[0022]

Figure 1 DETAILED DESCRIPTION OF THE INVENTION is an external view of the image processing apparatus 1 according to an embodiment of the present invention. Image processing device 1
Is a computer that performs correction by specifying the red-eye region in the image by executing the program. As shown in FIG. 1, the image processing apparatus 1 includes a keyboard 111 and a mouse 112 for receiving an input from a user, and a display 12 for displaying an instruction menu for the user, an acquired image, and the like.

The image processing apparatus 1 has a fixed disk 161 for storing data such as images inside, further,
A recording disk 91 storing a program and a memory card 92 storing image data can be loaded into a reading device 162 and a card slot 163 as computer-readable recording media, respectively.

FIG. 2 is a block diagram showing the configuration of the image processing apparatus 1. The image processing apparatus 1 includes a CPU 13, a RAM,
14 and a ROM 15 connected to a bus line. The bus line further includes a display 12, a keyboard 111 and a mouse 112 for receiving input from an operator, a fixed disk 161 for storing data and programs, and a recording disk 91 (optical disk, magnetic disk, magneto-optical disk, etc.). in reader 162 for transferring the information,
In addition, a card slot 163 that exchanges information with the memory card 92 is appropriately connected via an interface (I / F) or the like.

The RAM 14, fixed disk 161, reader 162, and card slot 163 can exchange data with each other.
The display 12 can display various information and images stored in the RAM 14, the fixed disk 161, the memory card 92, and the like.

The program 141 shown in FIG. 2, fixed disk 16 via the reading device 162 from the recording disk 91
Stored in 1, has been transferred from the fixed disk 161 to the RAM 14, is it possible to execute by CPU 13.

FIG. 3 is a diagram showing a functional configuration realized by the CPU 13 operating according to the program 141 in the RAM 14 together with other configurations. In the configuration shown in FIG. 3, the display control unit 201, the target area determination unit 202,
The size determining unit 203, the characteristic region extracting unit 204, the red-eye region specifying unit 205, and the correcting unit 206 indicate functions realized by the CPU 13 and the like.

The display controller 201 controls the display of an image on the display 12 based on the image data. The target area determination unit 202 receives from the user via the mouse 112 a determination of a target area to be processed in the image indicated by the image data 301. The size determination unit 203 determines the size of a unit area that is a unit for performing a process described below for the target area.

The characteristic region extracting section 204 performs a process for each unit region on the target region, and extracts a plurality of types of characteristic regions according to the characteristic amount. Note that saturation, lightness, and hue are used as the feature amounts. Eye region identification unit 205 identifies the characteristic region constituting the red region of the plurality of characteristic regions extracted. The correction unit 206 performs color conversion on the specified red-eye region, and corrects the red-eye in the image to a normal color eye. The corrected image data is the corrected image data 3
It is stored in the 02 as RAM14.

FIG. 4 is a block diagram showing a functional configuration of the correction unit 206. The correction unit 206 includes a histogram generation unit 211 that generates a histogram of a feature amount in a feature region forming a red-eye region, a correction curve generation unit 212 that generates a correction curve for correcting red eye, and a feature amount using the correction curve. conversion unit 213 for converting the filter generation unit 214 for generating a filter for smoothing, and a smoothing unit 215 performs smoothing of the red-eye region after correction. Details of these functions will be described later.

FIGS. 5 to 7 are diagrams showing the flow of operation when the image processing apparatus 1 specifies and corrects a red-eye area. Hereinafter, the operation of the image processing apparatus 1 for correcting red-eye (hereinafter, referred to as “red-eye correction”) will be described with reference to FIGS. 3 to 7.

First, when the user selects desired image data from the memory card 92 and the fixed disk 161 using the keyboard 111 and the mouse 112 while looking at the display 12, the selected image data 301
While being read by the AM 14, the display control unit 201 displays an image on the display 12 based on the image data 301 (step S11).

The user uses the mouse 1 to specify the area where red-eye correction is required.
12, the target area determination unit 20
2 determines a target area to be calculated (step S1
2). Specifically, the user as shown in FIG. 8 is an image 40
By designating two diagonal points of the target area 402 in one, the target area 402 including the red-eye area corresponding to one eye
Is determined.

When the target area 402 is determined, the size determining unit 203 determines the number of pixels on one side of the unit area (hereinafter referred to as “unit area size”) based on Equation 1 based on the number of pixels N1 and the parameter S of the target area 402. .) N (the size of the unit area is n × n) is obtained (step S13).

[0035]

(Equation 1)

[0036] The parameter S is a value that indicates the ratio of the number of pixels of the image 401 and the target region 402, is the number of pixels of the image 401 are parameters determined by the number 2 When N2.

[0037]

(Equation 2)

The function F is such that the smaller the parameter S is, the smaller the unit area size n to be output is.
The unit area size n to be output as the number N1 of pixels of pixel 2 is larger.
Is a function having an input / output relationship as shown in Equations 3 to 5, for example.

[0039]

(Equation 3)

[0040]

(Equation 4)

[0041]

(Equation 5)

[0042] By determining the unit area size by using such function F, the target area 4 to the image 401
02, the target area 402 can be divided into unit areas of a size corresponding to the ratio of the size of the target area 402, and thereafter, the calculation is performed in units of the unit area. A highly accurate red-eye area that is not easily affected can be specified. Note that the minimum value of n is 1, in which case one unit area is one pixel and the target area 40
2 will be processed for each pixel.

When the unit region size n is determined, the characteristic region extraction unit 204 determines the RG of each unit region of the target region 402.
The B value (average RGB value) is converted into a color space of the HSL color system, and the hue, saturation, and lightness of each unit area are obtained as feature amounts (step S14).

Next, the pupil region, the iris region, and the extra-iris region are extracted from the target region 402 as characteristic regions by the characteristic region extracting unit 204 based on the characteristic amount of the unit region. Specifically, a region where the saturation and brightness are within a certain range (a region where both saturation and brightness are low) is extracted as a pupil region, and a region where hue and saturation are within a certain range (hue is red). (A region in which the color is slightly yellow and the saturation is in a high range) is extracted as an iris region. Further, a region having the same lightness range as the pupil region and the same hue range as the iris region is extracted as an extra-iris region (step S15).

By extracting a characteristic region to be corrected based on the characteristic amount, it is possible to appropriately perform the correction based on the characteristic amount described later for each characteristic region.

[0046] Figure 9 is a characteristic region extracting unit 204 is a diagram illustrating the characteristic region extracted from the target region 402. Reference numeral 501 denotes the pupil area, codes 502a, 502b are iris region, reference numeral 503 denotes an iris area outside. The state of red-eye, various features region is sometimes plurality extracted Some If not extracted.

The range of the feature value used for extracting various feature regions is not determined from the feature value in the normal state where the red-eye effect does not occur, but is statistically determined from the feature value in the case where the red-eye effect occurs. Can be Therefore, it means that each feature region corresponding to such a feature amount has been extracted as a region where the red-eye effect may occur.
Conversely, a region that is not extracted as any of the characteristic regions is determined to be a region that does not require red-eye correction.

Next, unnecessary regions that are not suitable for red-eye correction are excluded from the extracted characteristic regions (step S16).
For example, among the characteristic regions shown in FIG.
02b is excluded from the correction target as an iris region that is erroneously extracted. Further, based on the inclusion relation and the degree of contact of each characteristic region, it is determined again whether each characteristic region is appropriate as a region to be corrected.

As a specific example, when the extra-iris region does not touch most of the surroundings of the iris region, the extra-iris region should not be corrected since the extra-iris region will give an unnatural eye if corrected. it is determined to be region. If there is a region other than the feature region between the pupil region and the iris region may receive the peripheral pupil region is not a red-eye, it is determined that the region where the pupil region should not be corrected. Appropriate correction is determined not to be correct any of the feature region because it is difficult state when the pupil region is not included fully in the iris region (in this case, red-eye correction is not executed) .

When at least one characteristic region constituting the red-eye region to be subjected to the red-eye correction is determined, information indicating the characteristic region and the characteristic amount are input from the red-eye region specifying unit 205 to the correction unit 206, and the red-eye region is specified. sequentially for each feature region constituting the region, processing is executed according to the red-eye correction. Note that the following processing is executed without depending on the type of the characteristic region.

First, the correction unit 206 determines one characteristic region included in the red-eye region as a processing target (hereinafter, referred to as a "target characteristic region") (step S21).
Then, the histogram generation unit 211 (see FIG. 4) counts the number of unit regions constituting the target feature region, and obtains the size of the target feature region as the count number of the unit region.

[0052] If the size of the target feature area is equal to or less than the predetermined value (step S22), and this fact is transmitted from the histogram generation unit 211 to the size determination unit 203, even smaller set size determination unit 203 to the unit area size By doing so, the resolution is increased, and the process returns to step S14 (step S23). This will increase the unit region number in the feature region is extracted again, it is possible to improve the reliability of parameters obtained from the histogram to be described later. As a result, the accuracy of red-eye correction can be improved.

When one unit area is one pixel, a process for increasing the resolution is performed on the target area 402 (or the feature area of interest) by the size determining unit 203. As a process for improving the resolution (that is, a process for enlarging an image), a linear interpolation method, a cubic convolution interpolation method, a B-spline interpolation method, or the like is used. As a result, the number of pixels of the feature region to be extracted again is increased.

Next, the histogram generation unit 211 determines a feature amount to be processed in the feature region of interest (one of saturation, lightness, and hue; hereinafter, referred to as “feature feature amount”) (FIG. 6). : Step S31), correction of the feature amount of interest (Steps S32 to S35), and smoothing of the feature amount of interest (Steps S42 and S43).

[0055] In the correction of the target feature amount is a histogram of the definitive target feature region noticeable feature quantity is generated by the histogram generation unit 211 (step S32). That is, a histogram showing a relationship between the value of the target feature amount and the number of unit regions having the value is generated.

When the histogram of the feature amount of interest in units of unit areas is generated, the correction curve generation unit 212 obtains a parameter that is a statistical value derived from the histogram (step S33), and uses the parameter to obtain each parameter. A correction curve indicating the relationship between the pre-correction value and the post-correction value of the feature amount of interest in the unit area is generated (step S34). Thereafter, the conversion unit 213 converts the feature amount of interest of each unit area using the correction curve (step S35). Thereby, the feature amount of interest in the feature region of interest is corrected.

The smoothing of the corrected feature amount is executed only when the feature amount is saturation or brightness (step S41), and the saturation or brightness variation emphasized by the correction (variation in a two-dimensional space). Is alleviated by smoothing. In the smoothing, a weighting coefficient for smoothing is obtained from the histogram (step S42), and the feature amount of interest in the unit area is corrected to a smoothed value (step S4).
3). Thereby, smoothing is performed to an appropriate degree for each characteristic region.

[0058] The next noticeable feature quantity when the correction and smoothing is completed for the target feature amount one is determined (step S4
4, S31), correction and smoothing are performed again. Incidentally, since the present embodiment smoothing for hue it is omitted, the saturation and lightness correction and smoothing is performed, only the correction for the hue is conducted. However, the hue may be smoothed.

Although the smoothing is performed after the correction in FIG. 6, the smoothing may be performed before the correction. in this case,
Steps S41 to S43 are replaced with steps S34 and S35. Variations in saturation and lightness to be emphasized is alleviated by the correction may perform smoothing prior to correction. Further, by performing the smoothing before the correction, the noise in the characteristic region can be reduced before the correction.

Next, the correction curve used for correction and the smoothing will be described in more detail.

FIG. 10 is a diagram exemplifying a correction curve 61 relating to saturation. In FIG. 10, the horizontal axis represents the saturation before correction (that is, the input of the conversion unit 213), and the vertical axis represents the saturation after correction (that is, the output of the conversion unit 213). Value XS
1 is the minimum value of the saturation before correction in the target feature area, the value XS3 is the maximum value of the previous saturation correction. Value XS2
Is the mode value of saturation (saturation of the most unit areas). These values are derived from the histogram of saturation.

In FIG. 10, the corrected values YS1, YS
2, YS3 respectively correspond to the values XS1, XS2, XS3. The value YS1, which is the minimum saturation, is made equal to the value XS1. Accordingly, the saturation of the red-eye region will be excessively reduced can be prevented in an image after correction.

The value YS3 is obtained by the function FS3 shown in Expression 6.

[0064]

(Equation 6)

[0065] function FS3 maximum value XS3 and, by the variance VS of the saturation obtained from the histogram of saturation as a parameter has the property to reduce the higher value YS3 dispersion VS is small. That is, the function FS3 is obtained by subtracting the larger value ZS3 from the maximum value XS3 as the variance VS becomes smaller, and obtaining
Find S3.

In the red-eye correction, it is preferable that the maximum value of the saturation is suppressed as small as possible. On the other hand, a problem that the atmosphere of interest, wherein the region and thus to reduce the maximum value of saturation when dispersed VS is greater is changed occurs.
Therefore, the saturation correction curve 61 has a characteristic of decreasing the maximum value of the saturation as the variance of the saturation is smaller by the function FS3.

Since the value YS3 is not allowed to be lower than the value YS1, the function FS3 also uses the value YS1 (ie, the minimum value XS1) as an auxiliary parameter.

On the other hand, the value YS2 is a function FS2
Required by The reason why the degree of decrease of the intermediate value of the saturation (meaning about the middle of the range of the saturation; the same applies hereinafter) is determined based on the mode XS2 is that the saturation near the mode XS2 is determined. Is greatly reduced, so that the purpose of red-eye correction of reducing the saturation can be easily and accurately realized.

[0069]

(Equation 7)

[0070] function FS2 is the mode XS2 and, to the dispersion VSM the mode near the histogram of saturation as a parameter, as the value YS2 dispersion VSM is small is a small value. That is, the function FS2 subtracts a large value ZS2 from the mode XS2 Smaller distributed VSM value Y
Find S2. Since the value YS2 must be a value between the values YS3 value YS1, the function FS3 value YS
3 (or the maximum value XS3 and the variance VS for obtaining the value YS3) and the value YS1 (ie, the minimum value XS
1) is used as an auxiliary parameter.

When the variance VSM near the mode is small, the saturation histogram is concentrated around the mode XS2. Therefore, by setting the value YS2 to a small value when the variance VSM is small, it is possible to appropriately obtain the effect of the red-eye correction of reducing the saturation. On the other hand, when the variance VSM is large, the histogram of saturation is not biased so much in the mode XS2, the effect can not be expected that even lowering the efficiency saturation value YS2 as a small value. Rather, the unevenness in saturation will be noticeable. Therefore, in the function FS2, as the variance VSM is smaller, a larger value is subtracted from the value XS2 to obtain a value YS2.

The value YS2 may be adjusted by the function FS2 based on the difference between the maximum value XS3 and the mode XS2. For example, when the difference between the maximum value XS3 and the mode XS2 is small, the value YS2 may be slightly increased so that the correction curve 61 is not excessively bent.

As described above, the values YS1, YS2, YS
When 3 is determined, the coordinates (XS1, YS1), (XS
2, YS2), 2 passing through three points (XS3, YS3)
A next curve is generated as a correction curve 61 relating to saturation.
It may be required correction curve 61 by other methods from the three points.

FIG. 11 is a diagram illustrating a correction curve 62 relating to lightness. The horizontal axis before correction lightness (i.e., the input of the conversion unit 213) in FIG. 11 is the vertical axis represents the brightness after correction (i.e., the output of the conversion unit 213). Value XL
1 is the minimum value of the brightness before correction, and the value XL3 is the maximum value of the brightness before correction. The value XL2 is the mode of lightness.
These values are derived from the brightness histogram.

In FIG. 11, similarly to FIG. 10, the corrected values YL1, YL2, YL3 are the values XL1, XL, respectively.
It corresponds to the L2, XL3. The value YL1, which is the minimum lightness after correction, is made equal to the value XL1, and the value YL3, which is the maximum lightness, is made equal to the value XL3. Value YL1 value XL
The reason for making the value equal to 1 is to prevent the red-eye region from being blackened in the corrected image and causing unnatural eyes. so-called catch-light in order to to those alive eye corrected to prevent the loss of (illumination area which is bright and is reflected) present in the.

[0076] value YL2 is determined by function FL2 shown by the number 8. The reason why the degree of decrease in the intermediate value of the lightness is determined based on the mode XL2 is that the lightness is greatly reduced in the vicinity of the mode XL2 as in the case of the saturation, so that the lightness is reduced. This is to easily and accurately realize the purpose of red-eye correction.

[0077]

(Equation 8)

The function FL2 uses the mode XL2 and the variance VL of the brightness histogram as parameters, and uses the variance VL
Is smaller, the value YS2 is smaller. That is,
Function FL2 determine the value YL2 by subtracting a larger value ZL2 from the mode XL2 higher variance VL is small. Note that the value Y
Since L2 needs to be equal to or greater than the value YL1, the function FL2
Also uses the value YL1 (ie, the minimum value XL1) as an auxiliary parameter.

[0079] If the variance VL is large, there are various lightness in the current feature region is assumed that there is a strong gradient regarding lightness. A typical example is
It is conceivable that a plurality of small regions having different lightness exist in the feature region of interest. Then, the value YL in such a case
When 2 results in a small value, the boundary of the different small regions of lightness in the current feature region after correction becomes clear, so that the unevenness of brightness is conspicuous.

Therefore, the function FL2 is a function that reduces the value YL2 only when the variance VL is small. As a result, it is possible to reduce the brightness as much as possible while preventing the occurrence of a pattern such as a step in the characteristic region due to the red-eye correction.

[0081] As in the case of saturation even function FL2, values YL2 based on the magnitude of the difference between the maximum value XL3 and the mode XL2 may be adjusted.

As described above, the values YL1, YL2, YL
When 3 is determined, the coordinates (XL1, YL1), (XL
2, YL2), 2 through three points (XL3, YL3)
The following curve is generated as a correction curve 62 relating to lightness.
As a result, a correction curve 62 that reduces the intermediate value while maintaining the minimum value and the maximum value of the brightness is obtained. of course,
The correction curve 62 may be obtained from the three points by another method.

FIG. 12 is a diagram illustrating a correction curve 63 relating to hue. The horizontal axis before correction hue (i.e., the input of the conversion unit 213) in FIG. 12 is the vertical axis represents the hue after correction (i.e., the output of the conversion unit 213). Value XH
Reference numeral 2 denotes a mode value of the hue histogram, and a range between the value XH1 and the value XH3 is a hue range that can be regarded as the same color as the hue indicated by the value XH2. However, the hue of the current feature within the region intended to be included within the scope between all values XH1 and values XH3. The corrected value in Fig. 12 YH1, Y
H2, YH3 corresponds to each value XH1, XH2, XH3.

The values XH1 and XH3 may be determined as values at one end and the other end of the hue range before correction in the feature region of interest.

The corrected values YH1, YH2, and YH3 are set in advance, and are set as values corresponding to a brown range for a brown iris and a blue range for a blue iris. Then, the coordinates (XH1, YH1), (XH2,
YH2), secondary curve passing through three points (XH3, YH3) is generated as a correction curve 63 relating to hue. Of course, the correction curve 63 may be obtained from the three points by another method.

By generating the correction curve 63 by the above-described method, the mode XH2 before correction is changed to the value YH
It is realized that the hue of the feature region of interest is corrected to a value within an appropriate hue range while being corrected to 2.

Next, describing calculation and smoothing of the weighting coefficients for smoothing performed on the chroma and lightness. Smoothing is a process of converting to a value obtained by weighted average with the feature quantity of eight unit areas around the features of the unit areas to be processed. For example, 3 × 3
If any of the (= 9) unit areas has a feature amount C _p (p indicates the position of the unit area) and a weighting coefficient for this unit area is a _p , the smoothed feature amount C _p number 9
Is required.

[0088]

(Equation 9)

FIG. 13 is a diagram showing weighting coefficients of a filter used for smoothing. As shown in FIG. 13, the weighting coefficient of the central unit area to be processed is set to 1, and the weighting coefficient of the surrounding unit area is set to 1 / W. However, the value W is a value of 1 or more, and when the feature amount of interest is saturation, the expression W
Thus, in the case of brightness, it is obtained by Expression 11.

[0090]

(Equation 10)

[0091]

[Equation 11]

The function FWS for saturation is the mode value XS2 of saturation.
And dispersing VS as a parameter, these parameters are determined from the histogram of saturation. Function F for brightness
WL also the mode XL2 and dispersion VL brightness parameters, these parameters are determined from the histogram of lightness.

[0093] As described above, when the variance of the feature amount is large, it is likely that there are a plurality of small regions having different feature amounts in the current feature region. In such a case, if the red-eye correction is performed, the difference between the saturation and the lightness of each small region becomes large, and unevenness occurs in the feature region of interest. Therefore,
Function FWS, FWL each dispersion VS, the degree of more VL is large smoothing determine the value W to be larger. Incidentally, the degree as the smoothing value W is close to 1 is increased.

When there are a plurality of small areas having different feature amounts in the feature area of interest, the mode X or saturation
The larger the S2 and XL2, the more pronounced the deviation of the saturation and lightness after correction. Therefore, the functions FWS and FWL determine the value W such that the larger the mode values XS2 and XL2, the greater the degree of smoothing. I do.

As described above, by determining the degree of smoothing based on the variance and the mode, appropriate smoothing according to the characteristics of each characteristic region is realized.

When the correction and the necessary smoothing for the feature region of interest are completed, it is confirmed whether or not an unprocessed feature region exists (FIG. 7: step S51). The next feature region of interest is determined (FIG. 5: step S21). Thereafter, correction and smoothing are performed on the saturation and brightness of each unit area of the feature region of interest, and correction is performed on the hue.

When the correction and smoothing are performed on all the characteristic amount regions, the target region 4 having the corrected characteristic region is obtained.
02 is returned from the HSL color space to the RGB color space, and further combined with the image data 301 to correct the corrected image data 3
It is stored in 02 as RAM 14 (step S5
2). The corrected image data 302 is transferred to the display control unit 201, and an image on which red-eye correction has been performed is displayed on the display 12 (step S53).

The image processing apparatus 1 for performing the red-eye correction according to the program 141 has been described above. The image processing apparatus 1 generates a histogram of the feature amount in the feature area to be corrected and corrects the correction curve for correcting the feature amount. Is generated based on the histogram. Therefore, it is possible to perform correction on the characteristic region according to the characteristics of the characteristic region. In addition, since a gradation occurs in the characteristic region after correction according to the gradation of the feature amount in the characteristic region before correction, correction can be made to a more natural eye. As a result, it is possible to appropriately correct red eyes that occur in various modes.

[0099] Incidentally, by using saturation, lightness and hue as the feature amount, can be easily performed based on the setting of the function for obtaining a correction curve to sensitive people, the more natural modified image Obtainable.

Also, after (or before) the correction of the characteristic region, the smoothing according to the characteristic region to be processed is performed, so that the entire red-eye region can be appropriately smoothed. Further, since the correction is performed after increasing the resolution for the small feature region, the correction can be appropriately performed even for the small feature region.

[0102] While there has been described one embodiment of the present invention, the present invention is susceptible to various modifications without being limited to the above embodiment.

[0102] Although in the above embodiment the correction for each feature region is performed, for example, these iris region when two iris region close is extracted may be treated as a feature region. The same applies to other types of characteristic regions.

[0103] Further, it is described that derive parameters from the histogram, the concept of a histogram showing the relationship between the feature quantity value and frequency may be expressed in any manner in the program. For example, a plurality of variables, arrays, or can be expressed histogram as an object.

For the hue, the values YH1, YH2, YH3
Has been described as being predetermined, but these values may be obtained using a function. For example, the values YH1, Y
H2 and YH3 may be obtained by a function using the hue range and mode before correction, the mode after correction, and the like as parameters.

In the above-described embodiment, an example will be described in which the maximum value, the mode value, and the variance of the feature amount of interest are used as parameters used for generating a correction curve and a filter for smoothing, and the minimum value is used as an auxiliary. but was, other parameters may be used. For example, an average value may be used instead of the mode value, or a median value (median) may be used. As described above, by using the statistical value derived from the histogram, it is possible to easily prepare parameters necessary for generating a correction curve and a filter for smoothing.

[0106] Further, the hue as the feature quantity in the above embodiment, although the saturation and lightness are ^{used, L * a * b *,} LU
A value in a color system (color space) such as V or XYZ may be used as a feature value, or an RGB value may be directly used as a feature value. Incidentally, correcting the characteristic amounts will eventually correspond to correct the pixel values.

For smoothing, another method different from the above embodiment may be used. For example, the filter may be 5 × 5 or more, and the weighted average may be performed only on the adjacent unit areas in four directions.

On the other hand, the acquisition of image data in the image processing apparatus 1 is not performed by reading the image data from the memory card 92 as in the above-described embodiment. May be obtained by transmitting and receiving signals to and from the device.
The image processing device 1 may be provided with an imaging unit to acquire image data.

The program 141 may be written in the ROM 15 in advance. Further, in the image processing apparatus 1, a series of image processing is all executed by software processing by the CPU, but a part or all of the processing can be realized by a dedicated circuit. In particular, rapid image processing is realized by executing the repetitive operation in the logic circuit.

In the above embodiment, the shape of the target area 402 is rectangular, but the present invention is not limited to this. For example, the shape may be an elliptical shape or a shape arbitrarily designated by a user. Further, in the above embodiment, the target area 40
2 is designated by the user, automatically target region 402 by the determination and the like by the image recognition or feature amount may be determined.

[0111]

According to the present invention, claims 1 to 9 and claim 14 are provided.
In the invention, the correction target region with respect to the correction suitable for the characteristic of the correction target region can be performed is realized is possible to appropriately correct the red-eye caused by various aspects.

According to the second aspect of the present invention, a correction target area suitable for correction can be extracted. According to the third aspect of the present invention, appropriate correction can be performed for each of a plurality of correction target areas.

According to the fourth aspect of the present invention, parameters used for generating a correction curve can be easily prepared.

[0114] In the invention of claim 5, it is possible to correct the appropriate saturation in accordance with the dispersion of the saturation, the invention of claim 6, leaving a so-called catch-light to the image after the correction it can.

[0115] In the invention of claim 7, it can be reduced efficiently saturation or lightness.

According to the eighth aspect of the present invention, the hue can be appropriately corrected.

Further, according to the ninth aspect of the present invention, correction can be performed based on saturation, brightness and hue.

[0118] In the invention of claim 10, 11 and 15,
By performing smoothing corresponding to the feature region, it is possible to perform appropriate smoothing for the entire eye area.

According to the eleventh aspect of the present invention, parameters for determining the degree of smoothing can be easily prepared.

According to the twelfth and sixteenth aspects, the accuracy of red-eye correction can be improved.

[Brief description of the drawings]

1 is a diagram showing an appearance of an image processing apparatus.

FIG. 2 is a block diagram illustrating a configuration of an image processing apparatus.

FIG. 3 is a block diagram illustrating a functional configuration of the image processing apparatus.

FIG. 4 is a block diagram illustrating a functional configuration of a correction unit.

FIG. 5 is a diagram illustrating a flow of an operation of the image processing apparatus.

FIG. 6 is a diagram illustrating a flow of an operation of the image processing apparatus.

FIG. 7 is a diagram illustrating a flow of an operation of the image processing apparatus.

FIG. 8 is a diagram exemplifying an image in which a target area is specified;

FIG. 9 is a diagram illustrating a characteristic region.

FIG. 10 is a diagram showing a correction curve relating to saturation.

FIG. 11 is a diagram illustrating a correction curve relating to brightness.

12 is a diagram showing a correction curve for hue.

FIG. 13 is a diagram illustrating weighting coefficients of a filter for smoothing.

[Explanation of symbols]

First image processing apparatus 61 to 63 correction curve 91 recorded disc 141 Program 501 pupil region 502a iris region 503 an iris area outside S15, S21~S23, S32~S35, S42, S
43 steps

────────────────────────────────────────────────── ─── of the front page continued (51) Int.Cl. ⁷ identification mark FI theme Court Bu (reference) H04N 1/46 H04N 1/46 Z 5L096 F-term (reference) 5B057 CA01 CA08 CA12 CA16 CB01 CB08 CB12 CB16 CC01 CE17 CH08 5C066 AA01 AA11 BA20 CA17 DD06 EA07 EB02 EC01 ED00 GA02 GA05 GA32 GA33 HA02 JA01 KA12 KC01 KD06 5C076 AA21 BA06 BB25 CB01 5C077 LL19 MP08 PP02 PP37 PP43 PP46 PP68 PQ08 PQ12 PQ18 PQ19 TT09 5C079 HB01 HB06 LA01 LA10 LA14 LA37 LB01 MA11 NA03 NA13 NA29 5L096 AA02 DA01 EA06 FA15 FA35 FA46

Claims (16)

[Claims] 1. A program for correcting red-eye in an image, wherein the computer-executed program causes the computer to generate a histogram of a feature amount in a correction target area included in a red-eye area; A step of generating a correction curve based on parameters derived from the following, and a step of correcting a feature amount in the correction target area according to the correction curve. 2. A program according to claim 1,
The program execution by the computer, the computer program characterized by further executing the step of extracting the correction target region from the image based on the feature quantity. 3. The program according to claim 2, wherein a plurality of correction target areas are extracted from the image, and the histogram is generated for each of the plurality of correction target areas. 4. The program according to claim 1, wherein the parameter includes any one of a minimum value, a maximum value, a mode value, an average value, and a variance of the feature amount. program which is characterized. 5. The program according to claim 4, wherein the feature quantity is saturation, and the correction curve has a characteristic that the maximum value of saturation decreases as the variance of saturation decreases. Features program. 6. The program according to claim 4, said feature amount lightness, wherein the correction curve has a characteristic of reducing an intermediate value of brightness while maintaining maximum brightness And the program. 7. The program according to claim 4, wherein the feature amount is saturation or lightness, and a degree of decrease of an intermediate value of the feature amount by the correction curve is a mode value of the feature amount. A program characterized by being determined on the basis of: 8. The program according to claim 4, wherein the characteristic amount is a hue, and the correction curve is obtained as a curve for converting a mode value of the hue into a predetermined value. Features program. 9. The program according to any one of claims 1 to 4, program the histogram, the saturation is a feature quantity, characterized in that it is determined for each of the brightness and hue. 10. A program for correcting the red eye in the image, is performed by a computer of the program,
A step of performing red-eye correction on a plurality of feature areas extracted as a red-eye area from an image based on the feature amount; and before or after the step of performing the red-eye correction, And a step of performing smoothing according to the characteristic region to be processed. 11. The program according to claim 10, wherein the execution of the program by the computer is:
The computer further comprising: generating a histogram of a feature amount for each of the plurality of feature regions; and determining the degree of smoothing based on a parameter derived from the histogram. Program to do. 12. A program for correcting the red eye in the image, is performed by a computer of the program,
The computer, and determining a correction target region in an image, and comparing the size and the predetermined value of the correction target region, when the size of the correction target region is smaller than the predetermined value, the A program for executing a step of increasing the resolution of a correction target area and a step of performing red-eye correction on the correction target area. 13. A computer-readable recording medium on which the program according to claim 1 is recorded. 14. A red-eye correction method for correcting red-eye in an image, comprising: generating a histogram of a feature amount in a correction target region included in the red-eye region; and generating a correction curve based on a parameter derived from the histogram. A red-eye correction method, comprising: generating; and correcting a feature amount in the correction target area according to the correction curve. 15. A red-eye correction method for correcting red-eye in an image, comprising the steps of: performing red-eye correction on a plurality of feature regions extracted from the image as red-eye regions based on a feature amount; red-eye correction method and performing before or after the step, and a step of performing corresponding smoothed in the characteristic region to be processed for each of said plurality of characteristic regions. 16. A red-eye correction method for correcting red-eye in an image, comprising the steps of: determining a correction target area in the image; comparing the size of the correction target area with a predetermined value; A step of increasing the resolution of the correction target region when the size of the region is smaller than the predetermined value; and performing a red-eye correction on the correction target region. .