WO2008018459A1

WO2008018459A1 - Image processing method, image processing apparatus, image processing program, and image pickup apparatus

Info

Publication number: WO2008018459A1
Application number: PCT/JP2007/065446
Authority: WO
Inventors: Akihiko Utsugi
Original assignee: Nikon Corporation
Priority date: 2006-08-08
Filing date: 2007-08-07
Publication date: 2008-02-14
Also published as: JP2009258770A

Abstract

An image processing method for determining whether an image is a particular type of image. The image processing method comprises acquiring an image consisting of a plurality of pixels; storing a look-up table indicative of a likelihood of a particular type of image for each pixel value and for each pixel position; generating, based on the acquired image, an image to be determined; obtaining, based on the pixel values and pixel positions of the pixels of the image to be determined, likelihoods of the particular type of image on those pixels by use of the look-up table; totalizing the obtained likelihoods of the particular type of image as to the pixels of the image to be determined; and determining, based on the totalization result, whether the input image is the particular type of image.

Description

Specification

Image processing method, image processing apparatus, image processing program, and imaging apparatus

Technical field

The present invention relates to an image processing method, an image processing device, an image processing program, and an imaging device that determine whether or not there is a specific image in an acquired image.

Background art

[0002] In digital image processing, there is a high demand for processing for detecting a medium facial image of a captured image. For example, in a digital camera, a process that converts a detected face area into a preferred color or gradation, a process that extracts the appearance of a specific person in a video image, or a suspicious person image that is extracted in a surveillance camera There is processing.

[0003] As image processing for detecting a face, an input image is reduced at various magnifications, face determination areas are set at various positions in the reduced image, and whether the face determination area is a face or not is determined. A method of determining by a face determination method is common. As conventional face determination methods, template matching, SVM (support vector machine), two Eurora network, AdaBoost, etc. have been proposed.

In face determination by template matching, face determination is performed by calculating the degree of matching between the template face image and the determination target image. In SVM face determination, by calculating the degree of coincidence between a large number of support vectors (template face images and non-face images selected from learning samples) and the image to be determined, it can be used for various nominations of the face image to be determined. It can respond flexibly.

[0005] The operation of face determination by a neural network (three-layer perceptron) is slightly similar to SVM, but uses multiple weighted coefficient maps obtained by learning instead of support vectors. The face determination operation by AdaBoost is a little like a neural network, but uses multiple rectangular filters selected by learning instead of the weighted coefficient map obtained by learning.

[0006] Patent Document 1: Japanese Patent Laid-Open No. 2005-44330

Disclosure of the invention Problems to be solved by the invention

[0007] However, face determination by template matching cannot flexibly cope with various nominations of the determination target face image. In particular, when the illumination conditions of the template face image and the discrimination target face image are different, there are many cases where the determination cannot be made correctly. In addition, face determination by SVM has a problem that enormous processing time is required to calculate the degree of coincidence between many support vectors and a determination target image.

[0008] Further, face detection by a neural network is faster than SVM, but has a problem that processing time required for determination is slightly longer than that of AdaBoost or the like. In addition, AdaBoost's face determination using a rectangular filter requires processing time because it needs to calculate the rectangular filter each time the force determination target area, which is faster than SVM or neural network, is changed.

Means for solving the problem

[0009] According to the first aspect of the present invention, the image processing method for determining whether or not the image is of a specific type obtains an image composed of a plurality of pixels, and determines the degree of image quality of the specific type as a pixel value and A lookup table shown for each pixel position is stored, a determination image is generated based on the acquired image, and the pixel value and pixel position of the pixel of the determination image are used to determine the pixel value of the pixel. The degree of likelihood of a specific type of image is obtained, the degree of likelihood of pixels of the obtained image for determination is integrated, and based on the result of integration, it is determined whether the input image is a specific type of image.

According to the second aspect of the present invention, in the image processing method according to the first aspect, the specific type of image is preferably a face image.

According to the third aspect of the present invention, in the image processing method according to the first or second aspect, the determination image is preferably generated by extracting an edge component of the acquired image.

According to the fourth aspect of the present invention, in the image processing method according to the first or second aspect, the determination image is an edge component having a concave structure in which the pixel value is recessed locally from the periphery of the acquired image. It is preferable to generate by extracting.

According to a fifth aspect of the present invention, in the image processing method according to the first aspect, the look-up table includes a pixel at a pixel position corresponding to a characteristic element of a specific type of image. If the edge component of the image is large, the degree of image-likeness of the specific type is set to a value larger than the degree of image-likeness of the specific type of image when the edge component is small. At the pixel position, the edge component of the pixel is large! /, The degree of image-likeness of a particular type of case! /, And the edge component is small! /, A value smaller than the degree of image-likeness of a particular type of It ’s better to do that.

According to the sixth aspect of the present invention, in the image processing method according to the second aspect, the determination image is generated by extracting an edge component of the acquired image, and corresponds to any region of the eye-nose mouth. At the position, if the edge component of the pixel is large, the degree of the image quality of the face is larger than the degree of the image quality of the face when the edge component is small, and it corresponds to the area other than the eyes and nose and mouth At the pixel position where the edge component of the pixel is large! /, The degree of the image quality of the face in the case is set to a small value compared to the degree of the image quality of the face in the case of the small edge component. Is preferred.

According to the seventh aspect of the present invention, in the image processing method according to any one of the first to sixth aspects, the lookup table includes a determination target image sample group belonging to a specific type of image and a specific type of image. It is preferably generated by statistical processing based on a non-determination target image sample group that does not belong.

According to the eighth aspect of the present invention, in the image processing method according to the seventh aspect, in the statistical process, the first process is performed based on the determination target image sample group by a process equivalent to that for generating the determination image. Generate an image sample group, generate a second image sample group based on the non-determination target image sample group, and the frequency P at which the pixel value at the pixel position (x, y) of the first image sample group is E (x, y) (E) and the image at pixel position (x, y) of the second image sample group.

1

The frequency P (x, y) (E) at which the prime value is E is obtained, and the degree V of the image-likeness of a specific type at that pixel with respect to the pixel value E at the pixel position (x, y) of the judgment image V ( x, y) is represented by V (x, y) = L (x, y) (E), and the lookup table L (x, y) (E) at the pixel position (x, y) is represented by L (x, y) y) (E) = f (P (x, y) (E), P (x, y) (E)), and the function ί (P (x, y) (E), P (x, y ) (E)) is assigned to P (x, y) (E)

1 2 1 2 1

It is preferably a monotonically increasing function in a broad sense, and substantially monotonically decreasing in a broad sense with respect to P ₂ ( _x , _y ) ( _E ).

According to a ninth aspect of the present invention, in the image processing method according to the eighth aspect, the function P (x, y) ( E), P ₂ (x, y) (E)) is f (P _i (x, y) (E), P ₂ ( _X , _y ) (E)) = log {ε)

}, And ε and ε are preferably predetermined constants.

1 2

According to a tenth aspect of the present invention, in the image processing method of the first aspect, a plurality of lookup tables corresponding to the degree of contrast are stored, the contrast of the acquired image is calculated, and the plurality of lookup tables are calculated. It is preferable to select a lookup table according to the contrast.

According to the first aspect of the present invention, the image processing method for determining whether or not the image is a specific type of image obtains an image composed of a plurality of pixels, and determines the degree of the image quality of the specific type as the pixel value and Stores a look-up table for each pixel position, generates multiple reduced images of images acquired at multiple different reduction magnifications, generates a judgment image based on multiple reduced images, and A determination target area is set for the first reduced image, which is one of the reduced images, and based on the pixel value of the pixel in the determination target area and the pixel position in the determination target area, the look-up table is used. The degree of image-likeness of a particular type of pixel is obtained, the degree of image-likeness of the particular type of pixel in the obtained determination target area is integrated, and an image corresponding to the determination target area in the acquired image is obtained based on the integration result.It is determined whether the image is of a specific type.

According to a twelfth aspect of the present invention, in the image processing method according to the first aspect, the second reduced image corresponding to the determination target region is further reduced with respect to the second reduced image further reduced than the first reduced image. The determination target area is further set, and a second lookup table is further stored for each pixel position corresponding to the pixel value and the second determination target area, indicating the degree of image-likeness of a specific type, and the second determination target Based on the pixel value of the pixel in the area and the pixel position in the second determination target area, the second look-up table is used to determine the degree of image-likeness of the specific type of the pixel, and the obtained second determination The degree of likelihood of a particular type of image of a pixel in the target area is integrated, and the result of integration of the degree of likelihood of a particular type of image in the pixel of the judgment target area and a particular type of image of the pixel in the second judgment target area The degree of uniqueness It is preferable to determine whether the image corresponding to the determination target area in the acquired image is a specific type of image based on the result of the integration.

According to a thirteenth aspect of the present invention, the image processing program is any one of the first to twelfth aspects. According to a fourteenth aspect of the present invention, the image processing apparatus is an image processing apparatus equipped with the image processing program of the thirteenth aspect. Is.

According to a fifteenth aspect of the present invention, the imaging apparatus is an imaging apparatus that mounts the image processing program according to the thirteenth aspect.

The invention's effect

[0010] Since the present invention is configured as described above, a specific type of image can be determined at high speed without being affected by various illumination conditions.

Brief Description of Drawings

FIG. 1 is a diagram showing an image processing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a flowchart of an image processing program executed by the personal computer 1.

FIG. 3 is a diagram showing an edge extraction target pixel and peripheral pixels with coordinates xy.

FIG. 4 is a diagram showing the result of creating a luminance concave image E (x, y) for various luminance structures.

[Figure 5] Four types of edge images E (x, y) to E (x, y) are generated for a specific facial luminance image

14

FIG.

[Fig.6] Creates a specific edge image! /, Generates facial appearance V (x, y), and calculates facial appearance V

It is a figure which shows the example which performed the process which issues 1SUM1.

[Fig.7] A specific value of the lookup table L (E) is shown for each edge size.

1 (x, y)

is there.

FIG. 8 is a diagram showing a flowchart of the processing after obtaining the face-likeness Vsum to Vsum of the partial image in the face determination processing in step S6 of FIG.

Four

FIG. 9 is a diagram showing a flowchart of a process for obtaining facial appearance L (E).

1 (x, y)

FIG. 10 is a diagram showing a flowchart of an image processing program according to a second embodiment executed by the personal computer 1.

[Figure 11] Normal image edge image and reduced size edge image It is a figure which shows the flowchart of the process after calculating | requiring face-likeness.

FIG. 12 is a diagram showing a configuration of a digital camera 100 that is an imaging apparatus.

BEST MODE FOR CARRYING OUT THE INVENTION

[0012] First Embodiment

FIG. 1 is a diagram showing an image processing apparatus according to an embodiment of the present invention. The image processing apparatus is realized by the personal computer 1. The personal computer 1 is connected to a digital camera 2, a recording medium 3 such as a CD-ROM, another computer 4, etc., and receives various images (image data). The personal computer 1 performs the image processing described below on the provided image. The computer 4 is connected via the Internet and other telecommunications lines 5.

[0013] The program executed by the personal computer 1 for image processing is similar to the configuration of FIG. 1, such as a recording medium such as a CD-ROM, other computer power via the Internet or other electric communication lines, and the like. Provided and installed in the personal computer 1. The personal computer 1 is composed of a CPU (not shown) and its peripheral circuits (not shown), and executes a program in which the CPU is installed.

[0014] When the program is provided via the Internet or another telecommunication line, the program is transmitted after being converted into a signal on a carrier wave carrying a telecommunication line, that is, a transmission medium. Thus, the program is supplied as a computer readable computer program product in various forms such as a recording medium and a carrier wave.

[0015] The personal computer 1 according to the present embodiment performs image processing for detecting a medium facial image of a photographed image. Specifically, an edge component is extracted based on the input image to generate an edge image, and it is determined whether there is a face image based on the generated edge image. The processing in the present embodiment is characterized by the edge component extraction method and the face determination method based on the edge image.

[0016] In the following, it is expressed that image processing is performed on an image, but actually it means that image processing is performed on input image data. Also, in this embodiment, an edge is a portion (area, pixel) where the luminance value or pixel value is smaller than the surrounding area, is larger than the surrounding area! /, And the value is protruding! / Locations (areas, pixels), steps This refers to the difference (area, pixel). In particular, the indented area (area, pixel) is called a concave edge, and the protruding area (area, pixel) is called a convex edge.

Hereinafter, image processing for detecting a medium facial image of an image captured by the personal computer 1 of the present embodiment will be described in detail. FIG. 2 is a diagram showing a flowchart of an image processing program executed by the personal computer 1.

[0018] In step S1, an image (image data) to be detected for a face photographed (captured) with a digital camera or the like is input (acquired). Each pixel of the input image includes R, G, and B color components, and each color component ranges from 0 to 255. In step S2, a luminance image Y is generated by the following formula based on R, G, and B of the input image. That is, the luminance image Y plane is generated.

Y = (R + 2G + B) / 4

In step S3, the generated luminance image is hierarchically reduced and output. For example, given by the reduction ratio _κ of 0.9 ⁿ for integer n of 0 to 3 1, and outputs the luminance image which has been reduced by the reduction magnification κ of the 32 patterns. As a reduction method, for example, Cubic scaling or linear scaling may be used. The reason why multiple reduced images are generated in this way is that it is unclear whether the input image has a face image of any size, so that it can handle face images of any size. .

[0020] In step S4, the reduced luminance image Y (x, y) force and four kinds of edge images E (x, y) to E (x, y) are generated by the following procedure. In the following, the x direction is the horizontal direction of the image

14

Or the horizontal direction and the y direction are vertical or vertical.

[0021] First, an image Y (x, y) smoothed in the vertical direction and an image smoothed in the horizontal direction are obtained from the following equations.

LV

Generate the image Y (x, y). To extract the edge component in the vertical direction, smooth the horizontal direction.

LH

This is because, in order to extract the edge component in the horizontal direction using the processed image data, it is preferable to use the image data in which the vertical direction is smoothed.

Y (x, y) = (Y (x, y- l) + 2 XY (x, y) + Y (x, y + 1)} / 4

LV

Y (x, y) = (Y (x- l, y) + 2 XY (x, y) + Y (x + l, y)} / 4

LH

[0022] Next, using the image Y (x, y) smoothed in the horizontal direction, the vertical direction

LH

Di-image E (x, y) is generated. Each pixel of the edge image is called an edge pixel. E '(x, y) = Min (Y (x, yl), Y (x, y + 2))

1 LH LH

-Min (Y (x, y), Y (x, y + l))

LH LH

E (x, y) = γ (E '(x, y))

Next, a vertical edge image E (x, y) is generated from the following equation.

2

E '(x, y) = I Y (x, y-l) -Y (x, y) |

2 LH LH

+ I Y (x, y + l) -Y (x, y) |

LH LH

E (x, y) = γ (E '(x, y))

twenty two

[0024] Next, using the image Y (x, y) smoothed in the vertical direction, the horizontal direction

LV

_C to generate the image E (x, y)

Three

E '(x, y) 2 Min (Υ (χ-1, y), Y (x + 2, y))

3 LV LV

Min (Y (x, y), Y (x + 1, y))

LV LV

E (x, y) = γ (E '(x, y))

3 3

Next, a lateral edge image E (x, y) is generated from the following equation.

Four

E '(x, y) = I Y (x-l, y) -Y (x, y) |

4 LV LV

+ I Y (x + l, y) -Y (x, y) |

LV LV

E (x, y) = γ (E '(x, y))

4 4

Here, Min () is a function that returns the minimum value of 0. Ί (E) is a function that performs γ conversion and clipping. It performs the following operations and outputs an integer from 0 to 31. This Ml N () process is a nonlinear filter process. It can also be called non-linear filter processing, including γ conversion and clipping processing.

When Ε <0 γ (Ε) = 0

Ε> 63 ( _Ί ) = 31

When 0≤Ε≤63 γ (E) = (int) (4Χ ^ Ε)

The generation of the edge image will be described in more detail with reference to FIG. FIG. 3 is a diagram in which the edge extraction target pixel and the surrounding pixels are represented by coordinates xy. The above E ′ (x, y) is the luminance image Y (x, y) plane, 4 pixels in the vertical direction Y (x, y— 1), Y (x, y), Y (x, y + l),

LH LH LH LH

Out of Y (x, y + 2), the outer two pixels Y (x, y— 1), Y (x, y) based on the target pixel ( _x , y)

LH LH LH

The difference between the minimum value of +2) and the minimum value of the inner two pixels Y (x, y) and Y (x, y + l) is obtained. [0028] The positive value of E ′ (x, y) indicates that the value in the vicinity of the target pixel (x, y) is smaller than the value in the vertical peripheral pixel, that is, the pixel value is in the vertical direction. Indicates that it is dented from the surroundings. Therefore, the value of E (x, y) generated in this way is treated as a pixel value, and the generated image is called a vertical luminance concave image.

[0029] The above E '(x, y) is adjacent to the target pixel (x, y) in the vertical direction in the luminance image Y (x, y) plane.

2 LH

A value obtained by adding a difference in luminance value with the adjacent pixel is shown. That is, a large value is generated when the luminance value changes greatly between the adjacent pixels in the vertical direction. Therefore, the value of E (x, y) generated in this way is treated as a pixel value, and the generated image is treated as a vertically adjacent pixel.

2

This is called a difference image. The vertically adjacent pixel difference image detects the edge of the concave structure, the edge of the convex structure, and the edge of the step without distinction.

[0030] The above E '(x, y) and E (x, y), E' (x, y) and E (x, y) are horizontal edge images

3 3 4 4

Is for generating. E '() selection £ (x, y), E' () selection £ (x,

For 1 1 2 2 y), the vertical and horizontal directions are reversed, and the rest is calculated in the same way. Therefore, E (x, y) generated in this way is the horizontal luminance concave image, and E (x, y) is the horizontal adjacent image.

3 4

This is called an elementary difference image.

[0031] FIG. 4 is a diagram showing the result of creating a luminance recess image E (x, y) for various luminance structures. Fig. 4 (a) shows a case where the luminance is concave, Fig. 4 (b) shows a case where the luminance is protruding, and Fig. 4 (c) shows a case where the luminance is stepped. As can be seen from Fig. 4, the luminance concave image has a positive value only when the luminance is concave. Therefore, if the negative value of the luminance recess image E ′ is clipped to 0, an edge image E (x, y) that reacts only to the luminance recess is generated.

[0032] According to the luminance concave image, the reaction is particularly good in a locally dark spot such as the eyes and nose and mouth.

Fig. 5 shows the four types of edge images E (x, y) to E (x, y

14

FIG. In fact, the luminance depression image has a sharp peak at the position of the eyes and nose! In particular, in the vertical luminance concave image E in FIG. 5, it reacts to the eyes, nostrils, mouth, etc. Among them, it reacts strongly to the eyes, nostrils, etc. and becomes white. In other words, the value of E at that position is large. Therefore, the face can be detected with high accuracy by analyzing such a luminance concave image. However, instead of using only the luminance concave image It is also desirable to use an edge image created by a conventional method.

[0033] The reason why the edge image is gamma-converted is to convert the edge amount into an appropriate feature amount E. In image analysis, a subtle difference in the edge amount in a place where there is almost no edge! / Has a larger meaning than a slight difference in the amount of edge in a part where there is a large edge! /. By applying gamma conversion to the edge amount, the above effect is realized, and there is almost no edge! / The difference in the edge amount at the location is converted into a large difference in the feature amount E. The difference in edge amount is converted into a small difference in feature amount E.

Next, returning to FIG. 2, in step S5, a face determination target area of 19 × 19 pixels is set for every other pixel of the reduced image, and a partial image of the edge image in that area is output. This is performed for all reduced images. The 19 x 19 pixel face detection target area is suitable for detecting the eyes, nose, mouth, etc. with about 2 pixels when the area is a face.

In step S 6, it is determined for each partial image of the edge image output in step 5 whether this area is a face image. In the present embodiment, the determination of the face image is performed by the method described below.

[0036] First, for each pixel position (x, y) (0≤x≤18, 0≤y≤18) of the partial image of the edge image E (x, y), the face of that position is based on the following equation: Produces the likelihood V (x, y). Facialness V (x, y) is a numerical expression of facialness at each pixel position, and indicates the degree and degree of facialness. V (x, y) may be said to be a likelihood representing a degree of likelihood as a face.

V (x, y) = L (E (x, y))

Here, L (E) is a look-up table created in advance for each pixel position (x, y) (0≤x≤18, 0≤v≤18) by statistical processing described later. position

, Y) represents the face likeness when the edge E (x, y) is E.

[0037] Then, the generated facial appearance V (x, y) is integrated for all pixels (x, y) (0≤x≤18, 0≤y≤18) to calculate facial appearance V.

FIG. 6 is a diagram illustrating an example in which the above processing is performed on a specific edge image. In the face-like image in Fig. 6, the part that looks like a face is displayed in white, and the part that does not look like a face is displayed in black. The face-like image generated from the face edge image shown in Fig. 6 (a) is large overall. Has a value. That is, the overall image is whitish. However, the facial image generated from the non-facial edge image shown in Fig. 6 (b) has small values in some places. That is, the image becomes dark in some places.

[0039] In the non-face example shown in Fig. 6 (b), the area corresponding to both sides of the eyes, the nose, and the mouth does not look like a face. In the face-like image, the pixel value of that area is a small value. Next black! Therefore, the value V obtained by integrating all the pixels of the face-like image of the non-face image is a small value.

SUM1

FIG. 7 shows specific values of the lookup table L (E) for each edge size.

1 (x, y)

FIG. In FIG. 7, the larger the face-like value, the more white it is displayed. In Fig. 7, the left side is the facial appearance when the edge is small, and the right side is the facial appearance when the edge is large.

. If all the values in the lookup table L (E) are illustrated,

1 (x, y)

Is generated from 0 to 31, so 32 diagrams from L (0) to L (31) can be created.

The In FIG. 7, only eight of them are shown for convenience of illustration.

[0041] Note that the look-up table L (E) in FIG. 7 has a specific value for each edge size.

1 (x, y)

It is the figure represented visually. Actually, a table of pixel values with the pixel position (X, y) as an argument is stored in the memory for each edge value. In other words, the table power S of the pixel value with 32 pixel positions (X, y) as an argument is stored in the S memory.

In FIG. 7, the diagram on the left represents the facial appearance when the edge is small. Looking at the figure on the left, the facial features of the eyes, nose and mouth are small. This means that if the edges of the eyes, nose, and mouth are small, the area is not likely to be a face. For example, in the non-face example in Fig. 6 (a), the edge of the part corresponding to the nose is small, so the part does not look like a face.

In addition, the diagram on the right side of FIG. 7 shows the facial appearance when the edge is large. Looking at the figure on the right, the face-likeness of parts other than the eyes, nose, and mouth is small. This means that if the edge of a part other than the eyes, nose, or mouth is large, that part does not look like a face. For example, in the non-face example shown in Fig. 6 (a), the edge of the part corresponding to the space between the eyes and both sides of the mouth is large, so the part is not likely to be a face.

[0044] That is, assuming that the face image is a specific type of image and the eyes, nose, mouth, etc. are characteristic elements of the specific type of image, it corresponds to the characteristic elements of the specific type of image. Pixel position In this case, the degree of likelihood of a specific type of image when the edge component of the pixel is large is set to a value larger than the degree of likelihood of the specific type of image when the edge component is small. In addition, at pixel positions corresponding to elements other than the characteristic elements of a specific type of image, the degree of a particular type of image when the edge component of that pixel is large is expressed as the degree of a particular type of image when the edge component is small V. As a small value compared to the degree of!

When the process of referring to the look-up table is arranged, first, in the partial image of the edge image E (x, y), the value of the edge of x = 0, = 0 is obtained. Next, look-up table L (E) corresponding to the value of edge E is selected from 32 look-up tables.

1 (x, y) 1

Decide. When the lookup table L (E) is determined, this lookup table L

1 (x, y) 1 1 vx.y)

The value of the pixel position (0, 0) of (E) is obtained. This is the faceness value at the pixel position (0, 0) of the edge image E (x, y). This process is sequentially performed until the pixel power of x = 0, y = 0, and the pixels of x = 18, y = 18, and the face-like image V (x, y) is obtained. Then, Vsum is obtained by summing all V (x, y).

[0046] Through the above processing, the facial appearance Vsum of the partial image is generated based on the edge image E (x, y). Then, based on the edge images E (x, y) to E (x, y), the facial image Vsu

twenty four

The process of generating m to Vsum is performed in the same way.

twenty four

[0047] FIG. 8 is a diagram showing a flowchart of the processing after obtaining the face-likeness Vsum to Vsum of the partial image in the face determination processing in step S6 of FIG. Face determination process in step S6

Four

In fact, as explained above, facial appearance Vsum to Vsum are generated step by step,

14

If the integrated evaluation value is larger than the threshold value, the face is determined. However, the process of comparing the evaluation value with the threshold value is performed at each stage as shown in Fig. 8, so that images that are clearly not faces are excluded at an early stage and at a stage so that efficient processing can be performed. RU

[0048] First, in step S11, an evaluation value for determining whether or not a partial image is a face image is set as the face likelihood Vsum of the edge image E (x, y). In step S12, it is determined whether or not the evaluation value is larger than a predetermined threshold value th. If this evaluation value is larger than the threshold value th, the process proceeds to step S13. If this evaluation value is not larger than the threshold value thl, the partial image is If it is not a face image, the face determination process for the target partial image is terminated.

[0049] In step S13, the evaluation value is changed to the evaluation value in step S11, and the face appearance of the edge image E (x, y) is determined.

2 Let Vsum be the value added. In step S14, this evaluation value is greater than a predetermined threshold th2.

2

If the evaluation value is larger than the threshold value th2, the process proceeds to step S15. If the evaluation value is not larger than the threshold value th2, the partial image is determined not to be a face image and the face determination of the target partial image is performed. End the process.

[0050] In step S15, the evaluation value is changed to the evaluation value in step S13, and the face appearance of the edge image E (x, y) is determined.

Three

Let Vsum be the value added. In step S16, this evaluation value is greater than a predetermined threshold th3.

Three

If the evaluation value is larger than the threshold th3, the process proceeds to step S17. If the evaluation value is not larger than the threshold th3, the partial image is determined not to be a face image and the face determination of the target partial image is performed. End the process.

[0051] In step S 17, the evaluation value is changed to the evaluation value in step S 15, and the face appearance of the edge image E (x, y) is displayed.

Four

Let Vsum be the value added. In step S18, this evaluation value is greater than a predetermined threshold th4.

Four

Judge whether or not. If the evaluation value is larger than the threshold th4 in step S18, it is finally determined that the partial image is a face image. If this evaluation value is not greater than the threshold th4, the partial image is not a face image, and the face determination process for the target partial image is terminated.

[0052] The partial image face determination process described above is performed for each partial image shifted by 1 bit in each reduced image, and all partial images that can be determined as face images are extracted. move on.

In step S 7, if it is determined in step 6 that a partial image is a face, the face size S and coordinates (X, Y) for the input image of the partial image are output. S 1, X, and Y are given by the following expression using the size S ′ = 19 of the face in the reduced image, the coordinates (Χ ′, Υ ′) of the face area and the reduction magnification κ.

S = S '/ κ

X = X '/ κ

Υ = Υ '/ κ

As described above, when a face image is included in the input image, the position and size of the face image are detected and output.

[0055] <Statistical processing> Next, the statistical processing described above will be described. That is, a method for obtaining the facial appearance L (E) of the pixel when the edge (x, y) of the pixel position (x, y) is E will be described. Figure 9

1 (x, y)

FIG. 10 is a diagram showing a flowchart of processing for obtaining the facial appearance L (E). This process

It is executed on the personal computer 1.

[0056] In step S21, images of several hundred or more faces are acquired. That is, several hundred or more faces are photographed (captured) with a digital camera or the like, and the images (image data) are acquired. The acquired image is an image composed of the same color components as the image input in step S1 of FIG. In step S22, the image of the face photographed is scaled so that the size of the face area becomes 19 × 19 pixels, and the partial images cut out of the face area are set as face image sample groups.

In step S23, several hundred patterns or more of 19 × 19 pixel non-face image sample groups are acquired. This is appropriately extracted from images other than the face photographed with a digital camera and made into a non-face image sample group. It is also possible to extract from an image showing a face while avoiding the face area. In this case, the user may appropriately designate the non-face image area from the image captured on the monitor.

In step S24, an edge component is extracted from the face image sample group to generate a face edge image sample group. This process is the same as the process for generating the edge image E (x, y) in the face detection process. In step S25, an edge component is extracted from the non-face image sample group to generate a non-face edge image sample group. This process is also performed in the same manner as the process for generating the edge image E (x, y) in the face detection process.

[0059] In step S26, the frequency P (x, y, E) at which the edge of (x, y) becomes E is obtained for the face edge image sample group! In other words, the image power with a pixel (X, y) value of E ^

Count. In step S27, for the non-face edge image sample group, (x, y

The frequency P (x, y, E) at which the edge of) becomes E is obtained.

Non-face

In step S28, the facial appearance L (E) of the pixel when the edge E (x, y) at the pixel position (x, y) is E is calculated by the following equation.

1 (x, y)

L (E) = log {(P (χ, γ, Ε) + 8) / (Ρ (χ, γ, Ε) + ε)}

1 (x, y) face 1 nonface 2

Where ε and ε are predetermined constants, introduced to suppress logarithmic divergence and overlearning.

1 2

is doing. The value of ε should be set to about 1/1000 of the average value of P (x, y, E). The value of ε should be set to several tens of times the value of ε.

twenty one

[0061] In the above equation (Ε), log {(P (χ, γ, Ε) + ε)} is a monotonically increasing function.

Kx, y) face 1

And log {l / (P (χ, γ, Ε) +8)} is a monotonically decreasing function. That is, facial appearance L Non-face 2

(Ε) shows an increase in the distribution of face image samples whose edge E (x, y) is E at the pixel position (x, y).

1 (x'y) 1

The value increases monotonously in the direction of addition, and the value decreases monotonously in the direction of increase in the distribution of non-face image samples whose edge E (x, y) is E at the pixel position (x, y). It is a function that does. The distribution of face image samples whose edge E (x, y) is E at the pixel position (x, y), and the non-face whose edge E (x, y) is E at the pixel position (x, y) The distribution of image samples is usually normal.

[0062] A look-up table L (E for converting the edge images E (x, y) to E (x, y) to facial appearance

2 4 2 (x, y)

) To L (E) to generate the edge component extraction process in steps S24 and S25 above.

4 (x, y)

In the same way as the processing for generating edge images E (x, y) to E (x, y) in face detection processing

twenty four

do it.

[0063] When the processing of the first embodiment described above is performed, the following effects are obtained.

(1) The positions of eyes, nose, mouth, etc. in the face image are locally darker than the surroundings. In the conventional edge extraction method, it was impossible to distinguish between a force having a locally dark structure, a force having a locally bright structure, or a force having another structure. However, by detecting the edge of the concave structure as described above and generating a concave image that is an edge image, it is possible to appropriately extract the eyes, nose, mouth, and the like that are locally dark structures of the face image. As a result, the face image can be accurately determined.

[0064] (2) According to the luminance concave image, it reacts particularly well to locally dark places such as the eyes and nose and mouth.

In fact, the luminance concave image has a sharp peak at the position of the eyes and nose. Therefore, the face can be detected with high accuracy by analyzing such a luminance concave image. In the present embodiment, the edge image created by the conventional method is used together with the luminance concave image alone, so that it is possible to determine the face with even higher accuracy! /. RU

Yes

(3) The reason why the edge edge is gamma-converted is to convert the edge amount into an appropriate feature amount E. In image analysis, subtle effects in areas with few edges The difference in the amount of edge has a larger meaning than the difference in the amount of edge slightly where there is a large edge! /. By applying gamma conversion to the edge amount, the difference in the edge amount at the point where there is almost no edge is converted into a large difference in the feature amount E, and the difference in the edge amount at the point where there is a large edge is the feature amount E. Is translated into a small difference. As a result, the difference in edge amount matches the difference in image structure. As a result, the accuracy of face determination is increased.

(4) As is clear from FIG. 4 of the above embodiment, it can be seen that the luminance concave image has a positive value only when the luminance is concave. Therefore, in this embodiment, the negative value of the luminance recess image E ′ is clipped to 0. As a result, an edge image E (x, y) that reacts only to the brightness dent is generated, and the processing using the edge image E becomes frustrating.

[0067] (5) A face image can be detected by a simple and high-speed process in which pixel values of an edge image are converted into facial appearance using a lookup table and integrated. Also, by determining the edge image, there is an effect of suppressing the influence of the lighting conditions when shooting the image.

Yes

[0068] Second Embodiment

In the second embodiment, a face determination method that is highly resistant to variations in the contrast of the determination target image will be described. The second embodiment is realized by the personal computer 1 as in the first embodiment. Therefore, the configuration of the image processing apparatus according to the second embodiment is referred to FIG. 1 of the first embodiment.

[0069] <Statistical processing>

First, statistical processing described below is performed to create a look-up table (LUT) for face determination according to the second embodiment. The creation of the lookup table of the second embodiment will be described below with reference to FIG. 9 of the first embodiment.

First, in the same manner as in Step S21 to Step S23 in FIG. 9, several hundred or more 19 × 19 pixel face image sample groups and several hundred or more 19 × 19 pixel non-face image sample groups are obtained. get.

Next, gain is applied to the pixel values of the face image sample group, and adjustment is performed so that the variance of the pixel values is about 100. Alternatively, a facial image sample group having a pixel value variance of less than 200 is extracted. The face image sample group adjusted or extracted in this way and the previously obtained face image sample group. Using the collected non-face image sample group, a look-up table for face determination is created in the same manner as in steps S24 to S28 in FIG. The look-up table obtained in this way is called a low-contrast face determination look-up table.

Next, a gain different from the above is applied to the pixel values of the face image sample group to adjust the pixel values to have a dispersion power of about 00. Alternatively, a face image sample group having a pixel value variance of 200 or more is extracted. Using the face image sample group adjusted or extracted in this way and the non-face image sample group obtained earlier, a look-up table for face determination is performed in the same manner as in steps S24 to S28 in FIG. Create The lookup table obtained in this way is called a high-contrast face determination lookup table.

Yes

[0073] Next, using the low contrast face determination look-up table and the high contrast face determination look-up table obtained as described above, image processing for detecting a face image from the captured image Will be described. FIG. 10 is a diagram illustrating a flowchart of the image processing program according to the second embodiment executed by the personal computer 1.

Steps S31 to S34 are the same as steps S1 to S4 in FIG. 2 of the first embodiment. In step S 38, an integral image I (x, y) of the luminance image and an integral image I (x, y) of the square of the pixel value of the luminance image are created based on the following expression! /.

Country

[Equation 1]

ζ · -0 0

In step S35, a face determination target region is set in the same manner as in step S5 in FIG. 2 of the first embodiment. In step S39, the distribution σ ² of the pixel values of the luminance image Y (x, y) in the face determination target region is calculated. Face detection target area with 4 points 0 ^) + \ ^), 0 ^ + 11) + \ ^ + 11) as vertices In this area, the value Ysum obtained by integrating the luminance image Y (x, y) and the value Ysum2 obtained by integrating the square of the luminance image are calculated by the following formula.

[Equation 2]

Ysum ^ I {x + w, y + h) -I (xl, y + h) -I (x + w, yl) + I (xl, yY) Ysuml = I ₂ (x + w, y + h) -I ₂ (xl, y + h) -I ₂ (x + w, yY) + Lix-l, yl)

[0076] According to the above calculation, the integral can be obtained simply by adding and subtracting the pixel values of the four points, so that high-speed calculation is possible. Then, the variance σ 2 of the pixel values of the luminance image Y (x, y) in the face determination target region is given by the following equation.

[Equation 3]

In step S40, when the variance σ is less than 200, a low-contrast face detection look-up table is selected. If the variance σ ² is 200 or more, a high-contrast face detection lookup table is selected. Incidentally, when the variance sigma ² is large indicates that the image of Koko contrast, when the variance sigma ² is small indicates that an image of low contrast.

[0078] In step S36, using the face detection lookup table selected in step S40, the face determination process is performed in the same manner as in step S6 of the first embodiment. _C

In 37, the detection result is output as in step S7 of the first embodiment.

[0079] According to the present embodiment, the contrast of the face determination target region is measured at high speed, and a face detection look-up table is selected according to the collation, thereby reducing various determination processing times. It is possible to make a highly accurate determination with respect to contrast. [0080] Note that the lookup table that differs depending on the contrast is used because an edge is excessively large in an image having a high contrast. In other words, a high-contrast face can be determined by using a high-level look-up table for those with high contrast.

[0081] Third Embodiment

In the third embodiment, a method for performing face determination with higher accuracy by using images of a plurality of different resolutions will be described. The third embodiment is realized by the personal computer 1 as in the first embodiment. Therefore, the configuration of the image processing apparatus according to the third embodiment is referred to FIG. 1 of the first embodiment.

[0082] <Statistical processing>

First, statistical processing described below is performed to create a look-up table (LUT) for face determination according to the third embodiment. The creation of the lookup table of the third embodiment will be described below with reference to FIG. 9 of the first embodiment.

First, a look-up table for face determination similar to that in the first embodiment is created in the same manner as in steps S21 to S28 in FIG. Hereinafter, this lookup table is referred to as a normal size face determination lookup table.

Next, the face image sample group acquired in the step S22 in FIG. 9 is reduced to a size of about 12 × 12 pixels. Similarly, the non-face image sample group acquired in step S23 in FIG. 9 is reduced to a size of about 12 × 12 pixels. Using the face image sample group and the non-face image sample group thus created, a look-up table for face determination is created in the same manner as in steps S24 to S28 in FIG. The lookup table obtained in this way is referred to as a reduced size face determination lookup table.

[0085] Next, using the normal size face determination lookup table and the reduced size face determination lookup table obtained as described above, a face image is detected from the captured images. Image processing will be described. Since the image processing program of the third embodiment executed by the personal computer 1 is the same as the flowchart of FIG. 2 of the first embodiment, the following description will be given with reference to FIG. To do.

[0086] Steps S1 to S4 are the same as Steps S1 to S4 in the first embodiment. It is like.

In step S5, a 19 × 19 pixel face determination target region is set every other pixel of the reduced image, and partial images of edge images E to E in that region are output. Etsu output here

14

This image is called a normal size edge image. Further, with respect to the reduced image, 0.9 for the ^four of the second reduced image reduced at reduced small magnification, the 19 X 19 pixels of the face determination target area 12 corresponding to the same object and the X 12 A reduced size face determination target area is set for pixels, and the partial images of edge images E to E created for the second reduced image in that area are set.

1 4 Outputs an image. The edge image output here is called a reduced size edge image.

[0088] In step S6, the likelihood of a face is calculated for the normal size edge image using the normal size face determination look-up table in the same manner as in the first embodiment. Further, the face-likeness is calculated for the reduced-size edge image in the same manner as in the first embodiment using the reduced-size face determination lookup table.

FIG. 11 is a diagram showing a flowchart of the processing after obtaining the facial appearance of each partial image with respect to the normal size edge image and the reduced size edge image as described above. In the face determination process in step S6, as in the first embodiment, the likelihood of a face is generated in stages, and if the integrated evaluation value is greater than the threshold value, the face is determined. However, by performing the process of comparing the evaluation value with the threshold value at each stage as shown in Fig. 11, it is clearly not a face! /, And the image is excluded at an early stage. Make it possible to perform efficient processing.

First, in step S 51, the evaluation value for determining whether or not the partial image is a face image is set as the face likelihood Vsum of the reduced size edge image E (x, y). In step S52, it is determined whether or not the evaluation value is larger than the predetermined threshold th. If the evaluation value is larger than the threshold th, the process proceeds to step S53. If the evaluation value is not larger than the threshold th, the partial image is a face image. If it is not an image! /, The face determination process for the target partial image is terminated.

[0091] In step S53, the evaluation value is reduced to the evaluation value in step S51, and the reduced size edge image E (x, y

2

) Face likeness Vsum is added. In step S54, the evaluation value is a predetermined threshold th.

2

If the evaluation value is greater than the threshold value th2, the process proceeds to step S55.If the evaluation value is not greater than the threshold value th2, the partial image is determined not to be a face image and the face determination of the target partial image is performed. The process ends. [0092] In step S55, the evaluation value is reduced to the evaluation value in step S53, and the reduced size edge image E (x, y

Three

) Face likeness Vsum is added. In step S56, this evaluation value is set to a predetermined threshold th.

Three

If the evaluation value is larger than the threshold th3, the process proceeds to step S57. If the evaluation value is not larger than the threshold th3, the partial image is not a face image and the face determination of the target partial image is performed. End the process.

[0093] In step S57, the evaluation value is converted into the evaluation value in step S55, and the reduced size edge image E (x, y

Four

) Face likeness Vsum is added. In step S58, the evaluation value is a predetermined threshold th.

Four

If the evaluation value is larger than the threshold th4, the process proceeds to step S59. If the evaluation value is not larger than the threshold th4, the partial image is determined not to be a face image and the face determination of the target partial image is performed. The process ends.

Step S59 force and step S66 perform the same processing for the normal size edge image. As a result, if the evaluation value is larger than the threshold th8 in step S66, it is finally determined that the partial image is a face image. If this evaluation value is not greater than the threshold th8, the partial image is determined not to be a face image, and the face determination process for the target partial image is terminated.

[0095] In step 7, as in the first embodiment, the face detection result is output.

[0096] According to the third embodiment, compared to the first embodiment, it is possible to perform a face determination process with higher accuracy by further evaluating the facialness of the reduced-size edge image. For example, if the face determination target area is 19 x 19 pixels, the eyes are about 2 pixels and easy to detect, but the mouth is about 4 pixels and difficult to detect. However, if the same face determination target area becomes 12 × 12 pixels, the mouth becomes about 2 pixels and it is easy to detect. Therefore, by adding the evaluation of the face likeness of the reduced-size edge image, it becomes easy to detect the concave structure such as the mouth, and the face determination process can be performed with higher accuracy.

[0097] Variation

In the above-described embodiment, an example of determining a face image has been described. However, the present invention can also be applied to images other than facial images. In other words, the present invention can also be applied to determining whether or not a certain type of image has a certain force in the acquired image. In such a case, a look-up table indicating the degree of image-likeness of the specific type for each pixel value and pixel position is prepared by statistical processing, and this look-up table is used for each pixel of the determination image. specific If you ask for the degree of image quality of the kind.

In the above embodiment, an example has been described in which a luminance concave image is generated as an edge image, and a locally dark spot such as the eyes and nose of the face is appropriately determined. However, the brightness of the mouth showing teeth and laughing, and the nose shining with light, are locally brighter than the surroundings. In order to appropriately detect such a locally bright spot on the face, a brightness convex image may be generated as an edge image by the following formula, and the facial appearance may be obtained in the same manner.

E '(x, y) = Max (Y (x, y-l), Y (x, y + 2))

5 LH LH

-Max (Y (x, y), Y (x, y + l))

LH LH

E (x, y) = γ (E '(x, y))

5 5

[0099] In the above-described embodiment, an example in which the face-likeness L (E) is calculated by the following equation has been described.

Yes

L (E) = log {(P (x, y, E) + ε) / (Ρ (x, y, E) + ε)}

1 (x, y) face 1 nonface 2

However, the following formula may be used.

L (E) = ^ {P (x, y, E)}-^ {P (x, y, E)}

1 (x, y) face non-face

The first term ^ {P (x, y, E)} is also a monotonically increasing function, and the second term ^ {P (x, y, E)} is also monotonically decreasing.

Non-face

It can be said to be a small function.

In the above embodiment, the example in which the personal computer 1 performs the image processing for detecting the face image from the captured image has been described. However, the above-described processing may be performed on the captured image in an imaging apparatus such as a digital still camera.

FIG. 12 is a diagram showing a configuration of a digital camera 100 that is such an imaging apparatus. The digital camera 100 includes a photographing lens 102, an image sensor 103 including a CCD, a control device 104 including a CPU and peripheral circuits, a memory 105, and the like.

The image sensor 103 captures (captures) the subject 101 via the photographing lens 102 and outputs the captured image data to the control device 104. The control device 104 performs image processing for detecting the face image described above on the image (image data) captured by the image sensor 103. Then, the control device 104 performs white balance adjustment and various other image processing on the image captured based on the detection result of the face image! /, And applies the image data after the image processing. Store in memory 105. Further, the control device 104 can also use the detection result of the face image for autofocus processing or the like. The image processing program executed by the control device 104 is stored in a ROM (not shown).

[0103] The processing described above can also be applied to a video camera. Furthermore, it can also be applied to surveillance cameras that monitor suspicious individuals and devices that identify individuals based on captured face images and estimate gender, age, and facial expressions. That is, the present invention can be applied to all devices such as an image processing device and an imaging device that extract and process a specific type of image such as a face image.

[0104] While various embodiments and modifications have been described above, the present invention is not limited to these contents. Other forms conceivable within the scope of the technical idea of the present invention are also included in the scope of the present invention.

[0105] The disclosure of the following priority application is incorporated herein by reference.

Japanese patent application No. 215943 (filed August 8, 2006)

Claims

The scope of the claims

[1] An image processing method for determining whether an image is of a specific type,

Acquire an image consisting of multiple pixels,

Stores a look-up table that shows the degree of image-likeness of a specific type for each pixel value and pixel position.

A determination image is generated based on the acquired image,

Based on the pixel value and the pixel position of the pixel of the image for determination, the look-up table is used to determine the degree of the image quality of the specific type of the pixel, and the image quality of the pixel of the determined image for determination is calculated. Accumulating the degree,

An image processing method for determining whether or not the input image is the specific type of image based on the accumulated result.

[2] In the image processing method according to claim 1,

The image processing method, wherein the specific type of image is a face image.

[3] In the image processing method according to claim 1 or 2,

The image processing method, wherein the determination image is generated by extracting an edge component of the acquired image.

[4] In the image processing method according to claim 1 or 2,

An image processing method in which the determination image is generated by extracting an edge component having a concave structure in which a pixel value is locally recessed from the periphery of the acquired image.

[5] The image processing method according to claim 1,

The lookup table is

At the pixel position corresponding to the characteristic element of the specific type of image, the edge component of the pixel is large! /, The degree of image quality of the specific type in the case! /, And the edge component is small! /, And a larger value than the degree of image quality of the specific type

At pixel positions corresponding to other than the characteristic elements of the specific type of image, the edge component of the pixel is large, the degree of image quality of the specific type in the case! /, When the edge component is small. An image processing method in which the value is smaller than the degree of image quality of the specific type.

[6] The image processing method according to claim 2,

The determination image is generated by extracting an edge component of the acquired image, and the face component in the case where the edge component of the pixel is large V at a pixel position corresponding to the area of the eye / nose! The degree of image-likeness! / Is set to a large value compared to the degree of image-likeness of the face when the edge component is small! /

At the pixel position corresponding to the region other than the eyes and nose and mouth, the edge component of the pixel is large! /, The degree of the image quality of the face when it is small, and the image of the face when the edge component is small! / An image processing method in which the value is smaller than the degree of thickness.

[7] Claims 1 to 6! / In the image processing method according to any one of claims! /,

The look-up table is an image processing method generated by statistical processing based on a determination target image sample group belonging to the specific type of image and a non-determination target image sample group not belonging to the specific type of image.

[8] In the image processing method according to claim 7,

In the statistical processing,

A first image sample group is generated based on the determination target image sample group by a process equivalent to the generation of the determination image, and the second image sample is generated based on the non-determination target image sample group! Generate image samples,

Frequency P (x, y) () where the pixel value at the pixel position (x, y) of the first image sample group is E

1

E) and the frequency P (X, y) (E) at which the pixel value at the pixel position (x, y) of the second image sample group becomes E,

For the pixel value E at the pixel position (x, y) of the image for determination, the degree V (x, y) of the particular kind of image at that pixel is expressed as V (x, y) = L (x, y ) (E), the lookup table L (x, y) (E) at the pixel position (x, y) given by L (x, y) (E) = f (P (x, y) (E ), P (x, y) (E))

1 2

And generate

The function ί (P (x, y) (E), P (x, y) (E)) is substantially monotonically increasing in a broad sense with respect to P (x, y) (E).

1 2 1

An image processing method that is an additive function and is a substantially monotonically decreasing function in the broad sense of P (x, y) (E).

[9] The image processing method according to claim 8, The function ί (P (x, y) (E)

1, P (x, y) (E)) is

2

ί (Ρ (x, y) (E), Ρ (x, y) (E)) = log {(P (x, y) (E) + ε) / (P (x, y) (E) + ε)},

1 2 1 1 2 2

The image processing method, wherein ε and ε are predetermined constants.

1 2

[10] The image processing method according to claim 1! /,

Stores multiple lookup tables according to the degree of contrast,

Calculating the contrast of the acquired image;

An image processing method for selecting a lookup table corresponding to the contrast from the plurality of lookup tables.

[11] An image processing method for determining whether an image is of a specific type,

Acquire an image consisting of multiple pixels,

Generating a plurality of reduced images of the acquired image with a plurality of different reduction magnifications; generating a determination image based on the plurality of reduced images;

A determination target region is set for a first reduced image that is one of the plurality of reduced images, and the lookup table is based on a pixel value of a pixel in the determination target region and a pixel position in the determination target region. Is used to determine the degree of image quality of the specific type at that pixel,

The degree of the image-likeness of the specific type of pixels of the obtained determination target region is integrated, and based on the integrated result, an image corresponding to the determination target region in the acquired image is the specific type image. An image processing method for determining whether or not there is.

[12] The image processing method according to claim 11!

A second determination target area corresponding to the determination target area is further set for the second reduced image further reduced than the first reduced image;

A second look-up table further indicating a degree of image-likeness of a specific type for each pixel position corresponding to a pixel value and the second determination target region;

Pixel values of the pixels in the second determination target area and images in the second determination target area Based on the raw position, the second look-up table is used to determine the degree of the particular kind of image at the pixel,

Integrating the degree of image quality of the specific type of pixels of the obtained second determination target region,

The acquired image based on the integration result of the degree of image specificity of the specific type of pixels in the determination target region and the integration result of the degree of image specificity of the specific type of pixels in the second determination target region. An image processing method for determining whether an image corresponding to the determination target region is an image of the specific type.

[13] An image processing program for causing a computer to execute the image processing method according to any one of claims 1 to 12.

14. An image processing apparatus equipped with the image processing program according to claim 13.

[15] An imaging device equipped with the image processing program according to claim 13.