JP6814172B2 - Skin internal structure estimation method, skin internal structure estimation program, and skin internal structure estimation device - Google Patents

Description

The present invention relates to methods, programs, and devices for estimating the internal structure of the skin. In particular, it relates to methods, programs, and devices for estimating the number density of capillaries in the papilla structure (papillary process) from images of the surface of the skin.

As shown in FIG. 17, the skin tissue (skin) S inside the human skin has a layered structure, and is divided into an epidermis 22, a dermis 24, and a subcutaneous tissue in order from the outermost layer. The epidermis 22 is composed of a membrane having a cuticle 20 in contact with the outer world as the outermost layer of the skin and a basal layer 26 in contact with the dermis 24, and the dermis 24 is a papillary layer 28 in contact with the basal layer 26 of the epidermis 22 and a subcutaneous tissue. It is composed of a film having a reticular layer 30 in contact with the dermis.
Further, between the epidermis 22 and the dermis 24, an uneven wavy papillary structure is formed in which the epidermis projection 27 of the basal layer 26 of the epidermis 22 and the papillary projection 29 of the papillary layer 28 of the dermis are present so as to bite into each other. Has been done.
Subcutaneous tissue contains arteries and veins that play important roles in skin nutrition, metabolism that carries waste products, and thermoregulation. The arteries and veins in the subcutaneous tissue are connected to the blood vessels (reticulated layer blood vessels) 32 in the reticular layer 30 of the dermis 24, and nourish the epidermis 22 in which no blood vessels exist via the capillaries 34 in the papillary process. And supplying oxygen.

In such a skin structure S, when the unevenness of the papillary structure is reduced and the surface area of the papillary process 28 is reduced due to aging or the like, the supply of nutrients and oxygen from the capillaries 34 inside the skin structure S to the epidermis 22 is delayed. Has been pointed out. As a result, the skin does not stretch and is said to gradually loosen. That is, the papillary structure and the capillaries within the papillary process are closely related to skin nutrition, metabolism, thermoregulation, and skin surface shape (eg, firmness, wrinkles, tarmi, etc.) and color (eg, dullness, etc.). It is believed to be involved.
Therefore, if the relationship between the state of the skin structure inside the skin and the state of the skin surface can be clarified, it is possible to evaluate the state of the skin and appropriately judge the change of the skin due to aging. As a result, the effects of cosmetics, skin care products, and skin care can be investigated, and can be applied to research or product development of cosmetics and the like.

As a method for clarifying the relationship between the state of the skin structure inside the skin and the state of the skin surface (or showing the relationship with the state of the skin surface), for example, Patent Document 1 describes scanning the skin surface. After acquiring the image and the microscopic image of the skin surface taken with a microscope and adjusting the scanned image so that the pores, pores, sweat glands, etc. can be clearly seen, the skin fineness such as the capillaries is collated with the microscopic image. A method of determining the microscopic field of view in a scanned image is shown using the structure as a marker. It has also been shown that changes in the number of blood vessels in capillaries can be easily grasped by observing specific capillaries at fixed points in microscopic observation using this method.
Patent Document 2 describes a method of providing a subject with a skin condition model (two-dimensional image, three-dimensional image, or model) that makes it easy to visually understand the analysis results of skin analysis including the skin surface to the internal structure of the skin. The skin information of the subject is acquired by imaging with a camera or a confocal laser microscope, the production of a skin replica, and the like, and the acquired skin information is obtained from at least one skin component (elements that characterize the skin surface structure [for example, an element that characterizes the skin surface structure]. Each skin component is compared with a pre-established visual score for [texture, sebum, pores, etc.] or elements that characterize the internal structure of the skin [including, for example, capillaries, papillary structure, and dermatitis, etc.]. Each characterizing parameter is calculated, and based on the calculated parameters, model parts that are part of the skin components associated with each score are selected, and the selected model parts are combined to make the skin condition model a subject. The method of providing is shown.
Patent Document 3 describes a method for clarifying the relationship between the state of the skin surface (skin "texture" and "skin color") and the internal structure of the skin (nipple height, number of nipples, and collagen-like structure). It has been proposed to use the estimation formula obtained by correlation analysis and regression analysis between the internal structure of the skin and the skin surface information.

In addition, in order to clarify the relationship between the state of the skin structure inside the skin and the state of the skin surface, it is necessary to properly analyze the state of the skin surface, but in particular, it is caused by aging etc. and gives an impression of appearance. Since wrinkles, which are an important factor that greatly influences the appearance of wrinkles, are an important factor that greatly influences the impression of appearance, a method for quantifying wrinkles easily and accurately from a skin image has been conventionally proposed. ..

For example, in Patent Document 4, a brightness image of the skin of a subject is acquired, a linear texture image at a predetermined angle corresponding to a region to be analyzed is extracted from the acquired brightness image of the skin, and the extracted linear shape is obtained. The directional strength or direction of the shape on polar coordinates obtained by extracting the linear component strength of a predetermined angle from the texture image and connecting a plurality of polar coordinate points indicating the correspondence between the predetermined angle and the linear component strength. It has been shown that the state of wrinkles at the analysis target site is quantified by calculating the combination of sexual intensity and area.
In Patent Document 5, a brightness image of the skin of a subject is acquired, a linear texture image at a predetermined angle corresponding to a region to be analyzed is extracted from the acquired brightness image of the skin, and the extracted linear texture image is obtained. The linear component intensity of a predetermined angle is extracted from, and the linear component intensity of a predetermined angle and the wrinkles on the face image obtained by the regression analysis of the extracted linear component intensity and a plurality of sample face images are obtained. It has been proposed to quantify the state of wrinkles at the analysis target site using a relational expression with the visual score.

Further, since the dullness of the skin is an important factor that greatly affects the beauty of the appearance of the skin, a method for quantifying the dullness easily and accurately from the skin image has been conventionally proposed.
For example, in Patent Document 6, skin color information is obtained from a subject's skin image, the distribution status of pigment components such as melanin components is obtained from the obtained color information, and the dullness of the skin is evaluated based on the distribution status. It has been proposed to obtain a color unevenness index.
In Patent Document 7, an evaluation target area is set in a photographed image of the skin surface, the intensity of the color element of each pixel in the evaluation target area and its histogram are acquired, and the frequency of the acquired histogram and the strength of the color element are acquired. A method for evaluating the dullness of the skin has been proposed based on the maximum and minimum values of the histogram and the strength of the color element.

JP-A-2016-214567 JP 2015-047422 JP 2011-101738 JP-A-2015-062569 Japanese Unexamined Patent Publication No. 2017-012384 JP 2014-213065 JP-A-2014-087641

The method described in Patent Document 1 enables accurate fixed-point observation of specific capillaries, but each time measurement is performed, the acquired skin scan image is compared with a microscopic image, and the skin fineness such as capillaries is collated. Since it is necessary to determine the microscopic field of view in the scanned image using the structure as a mark, there is a problem that it takes time to grasp the internal structure of the skin.
The method described in Patent Document 2 facilitates elements that characterize the internal structure of the skin (capillaries, papilla structure, dermis, etc.) and elements that characterize the surface properties of the skin (texture, sebum, pores, etc.). Although it can be quantified, there is a problem that it takes time and effort to measure the internal structure of the skin and the properties of the surface of the skin each time the skin condition is provided.
In addition, the method described in Patent Document 3 also has a simple and highly accurate state of the skin surface (skin "texture" and "skin color") and the internal structure of the skin (nipple height, number of nipples, and collagen-like structure). Although the relationship with the skin can be clarified, there is a problem that the internal structure of the skin cannot be estimated accurately when the property of the skin surface is "wrinkles".
Further, although Patent Documents 4 and 5 disclose that the properties generated on the skin surface of the analysis target site quantify "wrinkles" based on the skin image, it is possible to estimate the internal structure of the skin. Can not.
Further, although Patent Documents 6 and 7 disclose that the dullness of the skin surface is quantified based on the skin image, the structure inside the skin cannot be estimated.

Therefore, the present invention solves the above-mentioned problems and provides a method, a program, and an apparatus for estimating the internal structure of the skin easily, in a short time, and accurately from the skin properties generated on the surface of the skin. The purpose.

The method for estimating the internal structure of the skin according to the present invention acquires a first image of the surface of the skin, calculates a plurality of skin indexes showing predetermined properties of the surface of the skin from the first image, and obtains them in advance. The internal structure of the skin that calculates the number density of capillaries based on multiple skin indexes calculated using the estimation formula that estimates the number density of capillaries existing in the papillary process inside the skin. It is an estimation method of.
Further, it is preferable to calculate the skin comprehensive index from a plurality of skin indexes and calculate the number density of capillaries from the skin comprehensive index using an estimation formula.
In the estimation formula, the number densities of the second images of a plurality of different skins and the capillaries corresponding to the second images are acquired in advance, and a plurality of skin indexes are calculated from each of the plurality of second images. Multiple skin comprehensive indexes are calculated based on the obtained multiple skin indexes, and regression analysis is performed using the calculated number density of capillaries corresponding to each of the multiple skin comprehensive indexes and the plurality of skin comprehensive indexes. It is preferable to obtain by doing so.

The first image is an uneven image including information on the unevenness of the surface of the skin, the predetermined property is wrinkles, and the skin index is a wrinkle index showing the characteristics of wrinkles, which is an image of unevenness on the surface of the skin. Is obtained, and in the unevenness image, the uneven portion of the skin surface having a certain depth or a certain area or more is extracted as wrinkles, a plurality of wrinkle indexes are calculated based on the extracted wrinkles, and an estimation formula is used. Then, the number density of capillaries may be calculated based on the calculated wrinkle index.

The wrinkle index is preferably calculated based on at least two of the number of wrinkles, the length of wrinkles, the area of wrinkles, the depth of wrinkles, and the angle of wrinkles.
The wrinkle index calculated based on the length of wrinkles is a plurality of wrinkles calculated by calculating the long axis when the peripheral edge of each wrinkle is approximately elliptical as the length of each wrinkle for a plurality of wrinkles extracted from the uneven image. It is preferable that the wrinkle length is the average value or the maximum value.
The wrinkle index calculated based on the area of wrinkles is the area of a plurality of wrinkles calculated by calculating the number of pixels in the wrinkle area occupied by each wrinkle as the area of each wrinkle for a plurality of wrinkles extracted from the uneven image. Is preferably the average value or the maximum value of.

The wrinkle index calculated based on the depth of wrinkles is calculated by calculating the average value of the pixel values in the wrinkle region occupied by each wrinkle as the average depth of each wrinkle for each of a plurality of wrinkles extracted from the uneven image. , It is preferable that the average depth of each wrinkle calculated is the average value of a plurality of wrinkles.
The wrinkle index calculated based on the wrinkle angle is the angle at which the long axis of the multiple wrinkles extracted from the uneven image is the horizontal direction of the uneven image when the peripheral edge of each wrinkle is approximately elliptical. It is preferably the average value of the calculated and calculated angles of the plurality of wrinkles, or the standard deviation value.

The estimation formula estimates the number density of capillaries existing in the papillary process inside the skin corresponding to the wrinkle comprehensive index based on the wrinkle comprehensive index calculated from a plurality of wrinkle indexes.
Calculate the number density of capillaries existing in the papillary process inside the skin using an estimation formula from the wrinkle comprehensive index obtained based on the multiple wrinkle indexes calculated from the uneven image of the skin surface. Is preferable.
The estimation formula acquires an unevenness image containing information on the unevenness of a plurality of different skin surfaces, and in each of the unevenness images including information on the unevenness of a plurality of different skin surfaces, an uneven portion having a certain depth or a certain area or more. Is calculated as wrinkles, and based on the extracted wrinkles, a plurality of wrinkle indexes indicating the characteristics of wrinkles are calculated in each of the images including the unevenness information of the surface of a plurality of different skins, and the calculated wrinkle indexes are used as a plurality of calculated wrinkle indexes. Based on, the wrinkle comprehensive index is calculated for each of the images containing the unevenness information of the surface of the plurality of different skins, and the capillaries existing in the papillary processes inside the plurality of different skins corresponding to each of the plurality of different skin comprehensive indexes. It is preferably obtained by calculating the number density of the above and performing regression analysis using the plurality of wrinkle comprehensive indexes and the number densities of the capillaries corresponding to each of the plurality of wrinkle comprehensive indexes.

The wrinkle comprehensive index is preferably calculated by the linear sum of a plurality of wrinkle indexes.

The uneven image is preferably obtained by imaging the skin surface from a predetermined distance.
The uneven image is preferably obtained by imaging a replica of the skin surface.
Concavo-convex images can be obtained using a confocal microscope, a light cutting method, or a Gray code pattern light projection method.

The first image is an image of the surface of the skin, the predetermined property is dullness, and the skin index is a plurality of dullness indexes showing the characteristics of dullness, and the image of the surface of the skin is acquired to obtain the skin. A plurality of dullness indexes may be calculated from the image of the surface of the surface, and the number density of capillaries may be calculated based on the calculated dullness indexes using an estimation formula.

An analysis area is set in the image, at least one of the brightness, saturation, and hue angle is calculated from the analysis area, and the dullness index is calculated based on at least one of the calculated brightness, saturation, and hue angle. It is preferable to calculate.
The dullness index preferably includes at least one absolute value of lightness, saturation, and hue angle.
The dullness index is obtained by calculating at least one of lightness, saturation, and hue angle from the skin region excluding the analysis region, and acquiring the analysis region based on at least one of lightness, saturation, and hue angle. It is preferable to include the difference from the skin area.
The dullness index preferably includes a standard deviation obtained based on at least one of lightness, saturation, and hue angle calculated from the analysis area.
The dullness index preferably includes at least one of the number of stains and the area ratio of the stains obtained based on at least one of the brightness, saturation, and hue angle calculated from the analysis area.
The analysis area is preferably the eye area.

The estimation formula estimates the number density of capillaries corresponding to the dullness comprehensive index based on the dullness comprehensive index calculated from a plurality of dullness indexes, and can be used as a plurality of dullness indexes calculated from an image of the skin surface. It is preferable to calculate the number density of capillaries using an estimation formula from the dullness comprehensive index obtained based on the above.
The estimation formula acquires an image containing a plurality of different skin dullness, calculates a plurality of dullness indexes from an image containing a plurality of different skin dullness, and based on the calculated dullness index, a plurality of different skins. The total dullness index is calculated for each of the images containing the dullness information on the surface of the skin, and the number densities of the capillaries existing in the papillary processes inside multiple different skins corresponding to each of the multiple comprehensive dullness indexes are obtained. It is preferably obtained by regression analysis using a plurality of comprehensive dullness indexes and a number density of capillaries corresponding to each of the plurality of comprehensive dullness indexes.
The dullness comprehensive index is preferably calculated by the linear sum of a plurality of dullness indexes.

The skin internal structure estimation program of the present invention includes a step of acquiring a first image of the skin surface and a step of calculating a plurality of skin indexes indicating predetermined properties of the skin surface from the first image. , Calculate the number density of capillaries based on a plurality of calculated skin indexes using a pre-obtained estimation formula for estimating the number density of capillaries existing in the papillary process inside the skin. It is for letting the computer perform the steps.

The first image is an uneven image including information on the unevenness of the surface of the skin, the predetermined property is wrinkles, and the plurality of skin indexes are a plurality of wrinkle indexes indicating the characteristics of wrinkles, and the skin In the step of acquiring a surface unevenness image and in the unevenness image, the uneven portion of the skin surface having a certain depth or a certain area or more is extracted as wrinkles, and a plurality of wrinkle indexes are calculated based on the extracted wrinkles. This step may be used to cause a computer to perform a step of calculating the number density of capillaries based on a calculated wrinkle index using an estimation formula.

The first image is an image of the surface of the skin, the predetermined symptom is dullness, and the plurality of skin indexes are a plurality of dullness indexes indicating the characteristics of dullness, and an image of the surface of the skin is acquired. A computer with steps, a step of calculating a plurality of dullness indexes from an image of the surface of the skin, and a step of calculating the number density of capillaries based on a plurality of calculated dullness indexes using an estimation formula. It may be the one to be executed.

The device for estimating the internal structure of the skin of the present invention acquires an image acquisition unit that acquires a first image of the surface of the skin and a plurality of skin indexes indicating predetermined properties of the surface of the skin from the first image. Based on a plurality of skin indexes calculated using the skin index calculation unit and an estimation formula for estimating the number density of capillaries existing in the papillae inside the skin, which is obtained in advance, the capillaries It is provided with a capillary number density estimation unit for calculating the number density.

The image acquisition unit is a concavo-convex image acquisition unit that acquires a concavo-convex image including information on concavities and convexities on the surface of the skin as the first image, and the skin index calculation unit is a wrinkle feature as a predetermined property from the concavo-convex image. It is a wrinkle index calculation unit that calculates a plurality of wrinkle indexes indicating, and the capillary number density estimation unit calculates the number density of capillaries based on a plurality of calculated wrinkle indexes using an estimation formula. The wrinkle index calculation unit extracts the uneven portion of the skin surface having a certain depth or a certain area or more as wrinkles in the uneven image, and calculates a plurality of wrinkle indexes based on the extracted wrinkles. It may be a thing.

The image acquisition unit acquires an image of the surface of the skin as the first image, and the skin index calculation unit obtains a plurality of dullness indexes showing the characteristics of dullness as predetermined properties from the image of the skin surface. It is a dullness index calculation unit to be calculated, and the capillary number density estimation unit may calculate the number density of capillaries based on a plurality of calculated dullness indexes by using an estimation formula.

According to the present invention, the internal structure of the skin can be easily, in a short time, and accurately estimated from the skin properties generated on the surface of the skin.
From the wrinkles that occur on the surface of the skin, the number density of capillaries existing in the papillary process inside the skin can be easily, quickly, and accurately obtained.
From the dullness that occurs on the surface of the skin, the number density of capillaries existing in the papillary process inside the skin can be easily and quickly and accurately obtained.

It is a block diagram which shows the structure of the estimation apparatus of the internal structure of the skin which concerns on Embodiment 1 of this invention. It is a block diagram which shows the structure of the wrinkle index calculation part of FIG. It is a schematic diagram for demonstrating the calculation method of the wrinkle index by the wrinkle index calculation part of FIG. It is a figure which shows the flowchart of the estimation method of the internal structure of the skin executed by the apparatus which concerns on Embodiment 1. FIG. It is a graph which shows the correlation between a wrinkle index (wrinkle length) and a visual evaluation value (eye area). It is a graph which shows the correlation between a wrinkle index (wrinkle area) and a visual evaluation value (eye area). It is a graph which shows the correlation between a wrinkle index (wrinkle depth) and a visual evaluation value (eye area). It is a graph which shows the correlation between the wrinkle index (wrinkle angle) and the visual evaluation value (eye area). It is a graph which shows the correlation between the wrinkle quantitative value (wrinkle comprehensive index) and the visual evaluation value (eye area). It is a graph which shows the correlation between the wrinkle quantitative value (wrinkle comprehensive index) and the number density (eye area) of the capillaries in the papillary structure (papillary process). It is a block diagram which shows the structure of the estimation apparatus of the internal structure of the skin which concerns on Embodiment 2 of this invention. It is a block diagram which shows the structure of the dullness index calculation part of FIG. It is a figure which shows the area set on the face of a subject. It is a figure which shows the flowchart of the method of estimating the internal structure of the skin executed by the apparatus which concerns on Embodiment 2. It is a graph which shows the correlation between the dullness quantitative value (dullness comprehensive index) and the visual evaluation value (eye area). It is a graph which shows the correlation between the dullness quantitative value (dullness comprehensive index) and the number density (eye area) of the capillaries in the papillary structure (papillary process). It is a schematic diagram which shows the structure of the skin tissue.

Embodiment 1
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
FIG. 1 shows the configuration of the skin internal structure estimation device according to the first embodiment of the present invention. This device estimates the number density of capillaries existing in the papillary process inside the skin from the wrinkles that occur on the surface of the skin.
The device for estimating the internal structure of the skin of the first embodiment uses a photographed image of the unevenness of the skin surface of the subject to capture the internal structure of the skin of the subject (specifically, the capillaries existing in the papillary structure). The unevenness image acquisition unit 2 is provided with an unevenness image acquisition unit 2 that captures a replica (sample) 1 of the skin of the subject and acquires an unevenness image including information on the unevenness of the skin surface. , The wrinkle index calculation unit 3, the wrinkle comprehensive index calculation unit 4, the capillary number density estimation unit 5, and the display unit 6 are sequentially connected. Further, the database 7 is connected to the wrinkle index calculation unit 3, the wrinkle comprehensive index calculation unit 4, and the capillary number density estimation unit 5. Further, the control unit 8 is connected to the uneven image acquisition unit 2, the wrinkle index calculation unit 3, the wrinkle comprehensive index calculation unit 4, the capillary number density estimation unit 5, and the display unit 6, and the operation unit 9 is connected to the control unit 8. Has been done.
Here, the number density of capillaries is the number of capillaries observed in the observation area per unit area.

The uneven image acquisition unit 2 acquires a captured image from a camera that has captured a replica 1 of the subject's skin (eyes: a portion where the skin around the eyes is soft) from a predetermined distance (not shown).
The replica of the subject's skin to be photographed is created according to the "replica collection method" shown in the "wrinkle evaluation method guidance" created by the Japan Cosmetic Industry Association, and the replica agent is applied to the subject's skin (eyes). After applying and solidifying, the solidified replica agent is peeled off to prepare a skin replica to which the unevenness of the skin surface is transferred.

The skin portion on which the replica is produced is not limited to the eye area, and may be a portion where the capillaries existing in the papillary process can be easily observed using a device such as a microscope.
The uneven image acquisition unit 2 performs processing such as light intensity correction and noise removal on the captured image input from the camera, and acquires a three-dimensional image in which the unevenness of the skin surface can be determined.

The uneven image acquisition unit 2 can acquire a three-dimensional image by a method using a confocal microscope shown in JP-A-9-61720, for example, instead of the photographed image acquired by the camera. In this method, the sample surface is two-dimensionally scanned multiple times with a focused laser beam while changing the relative distance between the sample (skin replica) 1 and the objective lens of the microscope, and the reflected light from the sample surface is detected. The sample surface receives light from the device, detects the omnifocal image that maximizes the brightness of each pixel and the position information in the Z-axis direction that maximizes the brightness, and based on the omnifocal image and the position information in the Z direction. Form a three-dimensional image of.

Further, a three-dimensional image may be acquired by an optical cutting method. In this method, one slit light is applied to a measurement object (skin replica), the reflected light is received by an imaging sensor, and triangular distance measurement is performed to acquire three-dimensional distance data for one line of slit light. can do. Therefore, by scanning the entire measurement object with the slit light, a three-dimensional image of the entire measurement object can be acquired.

It is also possible to acquire a three-dimensional image by the Gray code pattern light projection method. In this method, a projector is placed on a camera, gray code pattern light is projected from the projector onto an object, the object is imaged by the camera in this state, and the three-dimensional observation target is based on the captured image. It is for acquiring an image.

The wrinkle index calculation unit 3 receives an unevenness image from the unevenness image acquisition unit 2, extracts an uneven portion having a certain depth or a certain area or more as wrinkles based on the unevenness image, and is based on the extracted wrinkles. Therefore, a plurality of wrinkle indexes indicating the characteristics of wrinkles are calculated.
As shown in FIG. 2, the wrinkle index calculation unit 3 includes a wrinkle extraction unit 10, a number calculation unit 11 connected to the wrinkle extraction unit 10, a length calculation unit 12, an area calculation unit 13, and a depth. It has a calculation unit 14 and an angle calculation unit 15, and calculates wrinkle indexes (number, length, area, depth, and angle).

As shown in FIG. 3, the wrinkle extraction unit 10 decomposes the unevenness image input from the unevenness image acquisition unit 2 into multiple resolutions, and then performs binarization processing on each of the obtained plurality of resolution images. Extract wrinkles. Specifically, among various shadows generated when the skin replica is irradiated with light, a threshold brightness is set in order to extract uneven portions having a certain depth or a certain area or more, for example. By converting the brightness above the threshold value to 255 (white) and less than 0 (black) and converting the brightness into binary values, only shadows having a density equal to or higher than the threshold value are extracted as wrinkles.
The setting of the threshold value can be determined according to the target skin replica.

For the image binarized by the wrinkle extraction unit 10, an image having the optimum resolution for calculating the wrinkle index is selected (in FIG. 3, an image of 27.68 [μm / px]), and the wrinkle extraction unit The images are output to the number calculation unit 11, the length calculation unit 12, the area calculation unit 13, the depth calculation unit 14, and the angle calculation unit 15, respectively, which are connected to the 10.
The image selected here (that is, the image used to calculate the wrinkle index (number, length, area, depth, and angle)) may be an image having a resolution different depending on each wrinkle index. Good.
Further, the resolution of the image selected here is preferably the same as the resolution of each image selected when calculating each wrinkle index used for creating the linear model (wrinkle comprehensive index model) described later.
The optimum image used for calculating each wrinkle index is automatically selected as an image satisfying a predetermined condition, but is not limited to this, and the operator selects the optimum image via the operation unit 9. You can also do it.

The number calculation unit 11 calculates the number of wrinkles by labeling the image obtained by binarizing the uneven image with a threshold value and counting the number of labels.

The length calculation unit 12 first calculates the long axis of the plurality of wrinkles extracted from the uneven image as the length of each wrinkle when the peripheral edge of each wrinkle is approximated by an ellipse, and then the calculated length of the plurality of wrinkles. The average value of the lengths or the maximum value of the lengths of a plurality of wrinkles is calculated as a wrinkle index (length).

The area calculation unit 13 first calculates the number of pixels in the wrinkle region occupied by each wrinkle as the area of each wrinkle for the plurality of wrinkles extracted from the uneven image, and then the average value of the calculated areas of the plurality of wrinkles. , Or the maximum value of the area of a plurality of wrinkles is calculated as a wrinkle index (area).

The depth calculation unit 14 first determines the average value of the pixel values in the wrinkle region occupied by each wrinkle for each of the wrinkles extracted from the uneven image for the plurality of wrinkles extracted from the uneven image, and the average depth of each wrinkle. Then, the average depth of each calculated wrinkle is calculated as a wrinkle index (depth).

The angle calculation unit 15 calculates, as the wrinkle angle, the angle at which the long axis forms the horizontal direction of the uneven image when the peripheral edge of each wrinkle is approximately elliptical for a plurality of wrinkles extracted from the uneven image, and then calculated. The average value of the angles of a plurality of wrinkles or the standard deviation value of the angles of a plurality of wrinkles is calculated as a wrinkle index (angle).

The wrinkle index calculation unit 3 includes a wrinkle index (number) calculated by the number calculation unit 11, a wrinkle index (length) calculated by the length calculation unit 12, a wrinkle index (area) calculated by the area calculation unit 13, and a depth. At least two or more wrinkle indexes of the wrinkle index (depth) calculated by the calculation unit 14 and the wrinkle index (angle) calculated by the angle calculation unit 15 are output to the wrinkle comprehensive index calculation unit 5.

The wrinkle comprehensive index calculation unit 4 uses a linear model consisting of a linear sum of a plurality of wrinkle indexes stored in advance in the database 7, and uses a number calculation unit 11, a length calculation unit 12, an area calculation unit 13, and a depth. A comprehensive index for wrinkles (wrinkle comprehensive index) is calculated from a plurality of wrinkle indexes calculated by the calculation unit 14 and the angle calculation unit 15. The calculated wrinkle comprehensive index is output to the capillary number density estimation unit 5.

The capillary number density estimation unit 5 is based on the estimation formula stored in advance in the database 7 and the wrinkle comprehensive index calculated by the wrinkle comprehensive index calculation unit 4, and the internal structure of the subject's skin (that is, within the papillary structure). (Number density of capillaries) is estimated.

The database 7 stores a linear model for use in the wrinkle comprehensive index calculation unit 4, that is, a linear model composed of a linear sum of a plurality of wrinkle indexes. Specifically, the wrinkle index calculation unit 3 calculates in advance the wrinkle index of each skin replica created from a plurality of samples (skin) having different skin properties, and the sum of each wrinkle index, and is created by multiple regression analysis. In addition, a linear model (wrinkle comprehensive index model) composed of linear sums of a plurality of wrinkle indexes is stored in the database 7.
Here, as the image used for calculating each wrinkle index used for creating the linear model, an image having a resolution different depending on each wrinkle index can be used. A plurality of wrinkle indexes are calculated from images having a plurality of resolutions as shown in FIG. 3, a sum of a plurality of wrinkle indexes is calculated by any combination, and a correlation is obtained among a plurality of models created by multiple regression analysis. It is preferable to store the highly characteristic model in the database 7 as a model used for estimating the internal structure of the skin.
Further, when estimating the internal structure of the skin of the subject, the wrinkle index calculation unit 3 used the resolution of the image used for calculating each wrinkle index to create the linear model stored in the database 7 in this way. It is preferable that the resolution is the same as that of each image used for calculating each wrinkle index.

In addition, the database 7 stores an estimation formula for use in the capillary number density estimation unit 5, that is, an estimation formula obtained from a regression analysis of the comprehensive index and the number density of capillaries existing in the papillary process. ing. Specifically, the number densities of capillaries existing in the papillae of a plurality of samples (eyes) having different skin properties used for creating the linear model described above are measured in advance with a microscope and Japanese Patent Application No. 2017-051264. The number of capillary hooves is extracted from the captured image captured by the apparatus shown in the above, and the number per area of the specific region is calculated to obtain the number. Next, the wrinkle comprehensive index value based on the skin replica created from the plurality of samples (eyes) described above and the estimation formula created by regression analysis of the number density of the corresponding capillaries are stored in the database 7. ..

The display unit 6 includes, for example, a display device such as an LCD (liquid crystal display), and exists in the papillary structure which is an estimation result of the internal structure of the skin estimated by the capillary number density estimation unit 5. Display the number density of capillaries.
The operation unit 9 is for the operator to input information, and can be formed from a keyboard, a mouse, a trackball, a touch panel, or the like.
The control unit 8 controls each part in the estimation device of the internal structure of the skin based on various command signals and the like input from the operation unit 9 by the operator.

The uneven image acquisition unit 2, the wrinkle index calculation unit 3, the wrinkle comprehensive index calculation unit 4, the capillary number density estimation unit 5, and the control unit 8 cause the CPU (Central Processing Unit) and the CPU to perform various processes. It is composed of operation programs for the purpose, but they may be composed of digital circuits.
Further, a memory can be connected to the CPU via a signal line such as a bus. For example, an uneven image generated by the uneven image acquisition unit 2, a wrinkle index calculated by the wrinkle calculation unit 3, and a wrinkle comprehensive index calculation unit. The wrinkle comprehensive index calculated in 4 and the result estimated by the capillary number density estimation unit 5 are stored in a memory, respectively, and the images and indexes stored in this memory and the estimation result are under the control of the control unit 8. Can be displayed on the display unit 6 with.

Next, the operation of the skin internal structure estimation device according to the first embodiment will be described with reference to FIG.
First, the captured image obtained by imaging the skin replica created from the eyes of the subject is input to the uneven image acquisition unit 2 and processed to acquire the uneven image (step S1), and the wrinkle index is calculated. It is input to the part 3. The wrinkle index input unit 3 extracts uneven portions having a certain depth or a certain area or more (step S2), and calculates a plurality of wrinkle indexes indicating the characteristics of the wrinkles for the extracted wrinkles (step S3). , Is input to the wrinkle comprehensive index calculation unit 4. The wrinkle comprehensive index calculation unit 4 calculates a wrinkle comprehensive index based on a plurality of wrinkle indexes calculated by the wrinkle index calculation unit 3 in advance using the wrinkle comprehensive index model stored in the database 7 (step S4). , Output to the capillary number density estimation unit 5. The capillary number density estimation unit 5 uses an estimation formula previously stored in the database 7 from a regression analysis of the total wrinkle index and the number density of capillaries existing in the papillary process to obtain wrinkles. The number density of capillaries existing in the papillary process is estimated based on the wrinkle comprehensive index calculated by the comprehensive index calculation unit 4 (step S5).

Embodiment 2
FIG. 11 shows the configuration of the skin internal structure estimation device according to the second embodiment of the present invention. This device estimates the number density of capillaries existing in the papillary process inside the skin from the dullness that occurs on the surface of the skin.
The device for estimating the internal structure of the skin of the second embodiment uses an image of the skin surface of the subject to capture the internal structure of the skin of the subject (specifically, the number density of capillaries existing in the papillary structure). ) Is estimated. This device includes an image acquisition unit 42, a dullness index calculation unit 43, a dullness comprehensive index calculation unit 44, a capillary number density estimation unit 45, a display unit 46, a database 47, a control unit 48, and an operation unit. 49 and.
The image acquisition unit 42 is sequentially connected to a dullness index calculation unit 43, a dullness comprehensive index calculation unit 44, a capillary number density estimation unit 45, and a display unit 46. A database 47 is connected to the dullness index calculation unit 43, the dullness comprehensive index calculation unit 44, and the capillary number density estimation unit 45. Further, a control unit 48 is connected to the image acquisition unit 42, the dullness index calculation unit 43, the dullness comprehensive index calculation unit 44, the capillary number density estimation unit 45, and the display unit 46, and the operation unit 49 is connected to the control unit 48. Has been done.
Here, the number density of capillaries is the number of capillaries per unit area observed in the region where the properties (dullness) of the skin are analyzed by image analysis.

The image acquisition unit 42 acquires a photographed image of the skin surface of the subject from the camera C in which the face F of the subject is photographed. The image acquired here has an RGB color space. The camera C may be any camera that can capture the face F of the subject, for example, a digital camera, a CCD camera, or the like, and a captured image taken by a mobile phone such as a smartphone can also be used.

The dullness index calculation unit 43 calculates the dullness index indicating the characteristics of dullness, and as shown in FIG. 12, the color space conversion unit 51, the absolute value calculation unit 52, the color difference calculation unit 53, the variation calculation unit 54, It also has a stain calculation unit 55.
The absolute value calculation unit 52, the color difference calculation unit 53, the variation calculation unit 54, and the stain calculation unit 55 are connected to the color space conversion unit 51.
The color space conversion unit 51 converts the captured image (RGB image) input from the image acquisition unit 42 into an L ^* a ^* b ^* color space after performing preprocessing such as light amount correction and noise removal. To generate a color space conversion image. Further, the color space conversion unit 53 includes an L ^* component (brightness) indicating brightness from the color space conversion image, and an a ^* component, b ^* component, C ^* component (saturation), and H ^* component (hue) indicating color. The angle) is extracted to generate each component image (that is, L ^* component image, a ^* component image, b ^* component image, C ^* component image, and H ^* component image). Each of the generated component images is input to the absolute value calculation unit 52, the color difference calculation unit 53, the variation calculation unit 54, and the stain calculation unit 55.

The absolute value calculation unit 52, as shown in FIG. 13 set, from the input component image (image of a subject's face F) from the color space converting section 51, the eyes prone color unevenness and stain as the analysis region R ₁ To do. Then, each color component from the analysis region _{R 1} is calculated ^{(L *} component, ^{a *} component, ^{b *} components, ^{C *} component and h component) average value of the intensity of the (color component values). Further, the absolute value of the calculated average value is calculated, and at least one of these values is used as a dullness index.

The color difference calculation unit 53, as shown in FIG. 13 setting, from each component image (image of a subject's face F) input from the color space converting section 51, the eyes prone color unevenness and stain as the analysis region R ₁ and, setting the entire skin of the face F of a subject except for the analysis region R ₁ as an analysis region R _2. The average value of the intensities of each color component (L ^* component, a ^* component, b ^* component, C ^* component, and h component) of each analysis region R ₁ and R ₂ is calculated. Further, the color difference ΔL between the analysis region R ₁ and the analysis region R _2, calculated .DELTA.a, [Delta] b, the value of ΔC and [Delta] H, and at least one dull index of these values.

Variation calculating unit 54, as shown in FIG. 13, from each component image (image of a subject's face F) input from the color space converting section 51, the eyes prone color unevenness and stain as the analysis region R ₁ Set. Each color component of the analysis region _{R 1} ^{(L *} component, ^{a *} component, ^{b *} components, ^{C *} component and h component) is calculated, to calculate these values of the standard deviation stdl, STDA, STDb, the value of stdC and stdH , At least one of these values is used as a dullness index. The standard deviation calculated here indicates the color variation.

Stain calculating unit 55, as shown in FIG. 13, set the color space conversion image input from the color space converting section 51 (image of a subject's face F), the eyes prone stain as the analysis region R _1, detecting a stain portion from the evaluation area R _1. Here, the spot portion is a local value of the L ^* component (brightness) indicating the brightness and the a ^* component, the b ^* component, the C ^* component (saturation), and the H ^* component (hue angle) indicating the color. For example, the size (maximum value or diameter) is larger than 2 mm and smaller than 50 mm.
Stain calculation unit 55, these values are detected from the analysis region R ₁ based on the stain portion size that varies locally, the area ratio of the stain portions in the analysis region R _1, the number of stains portion in the analysis region R ₁ Is calculated. At least one of the number of stains and the area ratio of stains is used as a dullness index.

For example, if you set the analysis region R ₁ with respect to L ^* component image, stain moiety, for example, can be detected by generating a Dog image (Difference of Gaussian image). Specifically, a Dog image having a different Gaussian size is generated from the L ^* component image. Generally, the stain is 2 mm to 10 mm in size and has a frequency of 0.05 cycle / mm to 0.25 cycle / mm. Therefore, the stain calculation unit 55 performs Dog image processing so that the component having the frequency band of the stain is extracted. Further, at the time of this Dog image processing, the shape of each component is calculated from the threshold-processed binarized image, and the component having a round shape and its circularity (4π × area) / peripheral length ² of ² mm to 10 mm is stained. It can be detected as a component.
Further, in the spot portion, after the above-mentioned Dog image processing is performed, the redness value (a ^* component value) and the yellowness value (b ^* component value) are further set to predetermined threshold values (Δa ^* component value and Δb ^*). It can also be detected by extracting a component having a component value smaller than 2.5).

The stain calculation unit 55 generates a dog image from the color space conversion unit 51 using not only the L ^* component image but also the color component images such as the a ^* component image and the b ^* component image, and in the same manner as described above. , Spots can also be detected.
Further, the stain calculation unit 55 extracts, for example, a component having an intensity equal to or lower than a predetermined threshold value from the L ^* component image without generating a dog image, and performs principal component analysis, independent component analysis, etc. on the extracted component. It is also possible to detect the spot portion by carrying out.
The stain calculation unit 55 calculates at least one total area of the stain portion detected in this way as a dullness index.

The dullness comprehensive index calculation unit 44 uses an absolute value calculation 52, a color difference calculation unit 53, a variation calculation unit 54, and a stain calculation unit using a linear model consisting of a linear sum of a plurality of dullness indexes stored in advance in the database 47. From the plurality of dullness indexes calculated in 55, a comprehensive index for dullness (dullness comprehensive index) is calculated. The calculated dullness comprehensive index is output to the capillary number density estimation unit 45.
The dullness index used to calculate the dullness comprehensive index value is not particularly limited as long as it is at least 2 or more, but (1) variation in redness component (stdA), area ratio of stains, and (2) variation in brightness (2) It is preferable to use stdL) and the area ratio of stains, or (3) variation in saturation (stdC) and area ratio of stains. This is because the dullness comprehensive index value calculated using these two dullness indexes can create a model having a high correlation with the visual evaluation value.

The reason why the combination of the above (1) variation in redness component (stdA) and the area ratio of spots is preferable is that the more dull the eye area is, the more uneven the color and the spots are, and the dull part of the skin. The redness becomes smaller, and the redness of the skin becomes stronger in the areas without dullness. Therefore, the variation in redness becomes large in the entire eye area including the dull part and the non-dull part. In addition, the more dull the area is, the larger the area ratio of the stains is because there are stains in the eye area. Therefore, when these two dullness indexes are used, the tendency that the dullness comprehensive index value becomes larger as the dullness exists in the eye area is modeled.
The reason why the combination of the above (2) brightness variation (stdL) and the area ratio of spots is preferable is that the skin becomes darker in the dull area and there is no dullness, similar to the variation in the redness component of the skin. The brightness of the skin becomes brighter in the part. Therefore, the variation in brightness becomes large in the entire eye area including the dull part and the non-dull part.
In addition, the reason why the combination of the above (3) saturation variation (stdC) and the area ratio of the stain is preferable is that the skin saturation becomes small in the dull part and the dullness is reduced, as in the case of the variation in the redness component of the skin. The saturation of the skin increases in the areas without it. Therefore, the variation in saturation becomes large in the entire eye area including the dull part and the non-dull part.

The capillary number density estimation unit 45 is located in the papillary structure existing inside the skin of the subject based on the estimation formula stored in advance in the database 47 and the dullness comprehensive index calculated by the dullness comprehensive index calculation unit 44. Estimate the number density of capillaries.

The database 47 stores a linear model for use in the dullness comprehensive index calculation unit 44, that is, a linear model composed of a linear sum of a plurality of dullness indexes. Specifically, the dullness index calculation unit 43 calculates in advance each dullness index and the sum of each dullness index from images of a plurality of samples (skin) having different skin properties, and a plurality of dullness indexes are created by multiple regression analysis. A linear model (dullness comprehensive index model) composed of the linear sum of the dullness indexes is stored in the database 47.
From each image, a plurality of dullness indexes are calculated, the sum of the plurality of dullness indexes is calculated by any combination, and among the plurality of models created by the multiple regression analysis, the model with high correlation is selected as the skin. As a model used for estimating the internal structure of the above, it is preferable to store it in the database 47.

In addition, the database 47 stores an estimation formula for use in the capillary number density estimation unit 45, that is, an estimation formula obtained from a regression analysis of the dullness comprehensive index and the number density of capillaries existing in the papillary process. Has been done. Specifically, the number densities of capillaries existing in the papillae of a plurality of samples (eyes) having different skin properties used for creating the linear model described above are measured in advance with a microscope and Japanese Patent Application No. 2017-051264. The number of capillary hooves is extracted from the captured image captured by the apparatus shown in the above, and the number per area of the specific region is calculated to obtain the number. Next, the dullness comprehensive index value based on the images of the plurality of samples (eyes) described above and the estimation formula created by regression analysis of the number densities of the corresponding capillaries are stored in the database 47.

The display unit 46 includes, for example, a display device such as an LCD (liquid crystal display), and exists in the papillary structure which is an estimation result of the internal structure of the skin estimated by the capillary number density estimation unit 45. Display the number density of capillaries.
The operation unit 49 is for the operator to input information, and can be formed from a keyboard, a mouse, a trackball, a touch panel, or the like.
The control unit 48 controls each part in the estimation device of the internal structure of the skin based on various command signals and the like input from the operation unit 49 by the operator.

The image acquisition unit 42, the dullness index calculation unit 43, the dullness comprehensive index calculation unit 44, the capillary number density estimation unit 45, the display unit 46, and the control unit 48 are a CPU (Central Processing Unit) and various processes on the CPU. It is composed of operation programs for performing the above, but they may be composed of digital circuits.
Further, a memory can be connected to the CPU via a signal line such as a bus. For example, a skin image generated by the image acquisition unit 42, a dullness index calculated by the dullness index calculation unit 43, and a dullness comprehensive index calculation unit. The dullness comprehensive index calculated by 44, the result estimated by the capillary number density estimation unit 45, etc. are stored in a memory, respectively, and the images and indexes stored in this memory, and the estimation result are under the control of the control unit 48. Can be displayed on the display unit 46 with.

Next, the operation of the skin internal structure estimation device according to the second embodiment will be described with reference to FIG.
First, the face F of the subject is photographed by the camera C to acquire a photographed image (RGB image) of the face (step S11), and as shown in FIG. 11, the face F is input from the camera C to the image acquisition unit 42. The input captured image is output to the color space conversion unit 51 of the dullness index calculation unit 43.

The captured image (RGB image) input to the color space conversion unit 51 is subjected to preprocessing such as light source correction and noise removal, and then the color space of the captured image is converted into the L ^* a ^* b ^* color space. A color space conversion image is generated. Then, the color space conversion unit 51 extracts each component indicating brightness and color from the color space conversion image, and each component image (that is, L ^* component image, a ^* component image, b ^* component image, C ^* component. Image and H ^* component image) are generated. Color space conversion unit 51, each component image generated absolute value calculator 52, the color difference calculation unit 53, and outputs the variation calculating unit 54 and the stain calculation unit 55, the calculation unit, in a predetermined analysis region R ₁ Calculate the dullness index. (Step S12).

The plurality of dullness indexes calculated in this way are input from the dullness index calculation unit 43 to the dullness comprehensive index calculation unit 44. The dullness comprehensive index calculation unit 44 calculates the dullness comprehensive index based on the plurality of dullness indexes calculated by the dullness index calculation unit 43 using the dullness comprehensive index model stored in the database 47 in advance (step S13). ), Output to the capillary number density estimation unit 45. The capillary number density estimation unit 45 uses an estimation formula previously stored in the database 47 from a regression analysis of the total dullness index and the number density of capillaries existing in the papillary process to obtain dullness. The number density of capillaries existing in the papillary process is estimated based on the dullness comprehensive index calculated by the comprehensive index calculation unit 44 (step S14).

An example is shown in which the capillary density of the skin of a subject is estimated using an estimation device for the internal structure of the skin.
Example 1
(1) Wrinkle index According to the criteria of the anti-wrinkle function evaluation test guideline (Journal of the Japan Council for Quality Health Care Vol.30, No.4, pp.316-332, 2006) prepared by the Japan Council for Quality Health Care, 20 people Subjects were randomly selected and the area around the eyes was the target area.
First, for 20 subjects, each wrinkle index (length, area, depth, and angle) is calculated using the skin internal structure estimation device of the first embodiment of the present application, and the above-mentioned Visual evaluation was performed by 8 levels of wrinkle grade classification according to the anti-wrinkle function evaluation test guideline.
<8 levels of wrinkle grade classification>
0 No wrinkles 1 Slightly indistinct shallow wrinkles 2 Slightly clear shallow wrinkles 3 Clear shallow wrinkles 4 Slightly unpleasant wrinkles in clear shallow wrinkles 5 Slightly deep wrinkles are observed 6 Clear deep wrinkles are observed 7 Remarkably unpleasant wrinkles are observed

In FIGS. 5 to 8, the values of each wrinkle index (length, area, depth, and angle) are plotted with "◆" with respect to the visual evaluation value. The correlation between each wrinkle index and the visual evaluation value is R ² (correlation coefficient) = 0.7292, and the correlation between the wrinkle index (length) shown in FIG. 5 and the visual evaluation value is R ² (correlation coefficient) = 0.7292, and the wrinkles shown in FIG. The correlation between the index (area) and the visual evaluation value is R ² = 0.4683, the correlation between the wrinkle index (depth) shown in FIG. 7 and the visual evaluation value R ² = 0.209, and the wrinkle shown in FIG. The correlation between the index (angle) and the visual evaluation value was R ² = 0.5123. From this, it can be seen that each wrinkle index has a correlation with the visual evaluation value.

(2) Wrinkle comprehensive index model In addition, 11 subjects for randomly creating (A) wrinkle comprehensive index model (linear sum of a plurality of wrinkle indexes) from 20 subjects who performed the above visual evaluation values. Divided into 9 people and (B) 9 people to confirm the reliability of the created wrinkle comprehensive index model, and (A) multiple wrinkle indexes (length and angle) obtained from 11 people in the group and their From the sum, a wrinkle comprehensive index model (linear model) consisting of linear sums of multiple wrinkle indexes was created by multiple regression analysis, and multiple wrinkle indexes (length and angle) obtained from 9 people in group (B). Was substituted into the wrinkle comprehensive index model obtained from the (A) group to obtain the wrinkle comprehensive index.
In FIG. 9, the sum of the wrinkle indexes obtained from the 11 people in the group (A) is plotted with “■” against the visual evaluation value, and the wrinkle comprehensive index obtained from the 9 people in the group (B) is the visual evaluation value. Was plotted with "◇".
The wrinkle comprehensive index model is obtained by the formula y = 0.002906746 × x1 + 0.011803113 × x2-1.076706106, where y is the wrinkle comprehensive index, x1 is the wrinkle index (length), and x2 is the wrinkle index (angle). Was done.
As a result of obtaining the correlation between the value of the wrinkle comprehensive index and the visual evaluation value, the correlation coefficient R ² was 0.832. From this, it can be seen that a high correlation can be obtained between the wrinkle comprehensive index calculated based on the plurality of wrinkle indexes and the visual evaluation value.

(3) Estimate formula In accordance with the above-mentioned anti-wrinkle function evaluation test guideline, in addition to the 20 people selected to create the above-mentioned comprehensive wrinkle index model and confirm its reliability, the target part is the eye area, and further mentioned above. According to the criteria of the anti-wrinkle function evaluation test guideline, 14 subjects were randomly selected, and a plurality of wrinkle indexes (length and angle) were calculated using the skin internal structure estimation device of the present application. In addition, the wrinkle comprehensive index was calculated using the wrinkle comprehensive index model created by acquiring skin (eye) images from the above-mentioned 11 people from the acquired multiple wrinkle indexes.
In addition, the number of capillaries existing in the papillary structure was extracted from the images of the skin (eyes) of 34 subjects taken by the apparatus described in Japanese Patent Application No. 2017-051264, and the number density in a predetermined area 10 mm square was obtained. Calculated.
The estimated formula y = -6.658x + 22.703 was obtained by regression analysis of the calculated total wrinkle index and the number density of capillaries.

FIG. 10 is a plot of the wrinkle comprehensive index calculated from 34 subjects against the calculated number density of capillaries. Comparing the wrinkle comprehensive index with the number density of capillaries, the correlation function R ² was 0.2216 and the p-value was 0.004957.
In the present art, when the p-value is less than 0.05, it is recognized that there is a significant difference, and therefore, it was found that the above results are sufficiently correlated.

Example 2
(1) Dullness index Eighteen subjects with different skin properties (especially dullness) were randomly selected, and the eyes were targeted. For 18 subjects, the dullness index (variation of redness component a ^* : stdA and area ratio of stains) was calculated using the skin internal structure estimation device of the second embodiment, and is shown below. Visual evaluation was performed in 5 stages of classification.
<5 levels of category classification>
1 I don't feel dullness at all 2 I feel dullness but I don't mind 3 I feel a little dullness 4 I feel dullness 5 I feel a lot of dullness

(2) Dullness Comprehensive Index Model In addition, a dullness comprehensive index model (linear model) consisting of linear sums of a plurality of dullness indexes was created by multiple regression analysis from the sum of the plurality of dullness indexes obtained above.
In the dullness comprehensive index model, if y is the dullness comprehensive index, x1 is the dullness index 1 (variation of redness component a ^* : stdA), and x2 is the dullness index 2 (stain area ratio), y = 0.479556966 × x1 + 0.176530059 × It was obtained by the formula x2 + 1.493311663.

Further, as shown in FIG. 15, the sum of the obtained dullness indexes was plotted against the visual evaluation value. Result of the correlation between the value and the visual evaluation value of dullness general index, the correlation coefficient R ² was 0.500. From this, it can be seen that a high correlation can be obtained between the dullness comprehensive index and the visual evaluation value.
Further, from FIG. 15, as a stain index, the dullness comprehensive index model created by using the variation of the redness component and the area ratio of the stain tends to have a larger dullness comprehensive index value as the dullness exists in the eye area. It can be said that it is modeled.

(3) Estimating formula In addition to the 18 subjects described above, 16 subjects were randomly selected, and using the skin internal structure estimation device of the present application, a plurality of dullness indexes (variation of redness components and spots) were used. Area ratio) was calculated and visual evaluation was performed by the above method. In addition, the dullness comprehensive index was calculated using the dullness comprehensive index model created by acquiring skin (eye) images from the above-mentioned 18 persons from the acquired dullness indexes of 34 people.
In addition, the number of capillaries existing in the papillary structure was extracted from the images of the eyes of 34 subjects taken by the apparatus described in Japanese Patent Application No. 2017-051264, and the number density in a predetermined area 10 mm square was calculated.
The estimated formula y = -0.027x + 3.799 was obtained by regression analysis of the calculated dullness comprehensive index and the number density of capillaries.

FIG. 16 is a plot of the dullness comprehensive index calculated from 34 subjects against the calculated number density of capillaries. Comparing the dullness comprehensive index with the number density of capillaries, the correlation function R ² was 0.141 and the p-value was 0.029.
In the present art, when the p value is less than 0.05, it is recognized that there is a significant difference, and therefore, it was found that the above results are sufficiently correlated.

1 Skin replica (sample)
2 Concavo-convex image acquisition unit 3 Wrinkle index calculation unit 4 Wrinkle comprehensive index calculation unit 5, 45 Capillary blood vessel number density estimation unit 6, 46 Display unit 7, 47 Database 8, 48 Control unit 9, 49 Operation unit 10 Wrinkle extraction unit 11 Number Calculation unit 12 Length calculation unit 13 Area calculation unit 14 Depth calculation unit 15 Angle calculation unit 20 Square layer 22 Epidermis 24 Dermis 26 Base layer 27 Epidermis protrusion 28 Papillary layer 29 Papillary protrusion 30 Reticulated layer 32 Blood vessels in the reticulated layer (reticular) Layer blood vessels)
34 Capillaries in the papillary process (straps)
41 Sample (face surface of the subject)
42 Image acquisition unit 43 Dullness index calculation unit 44 Dullness comprehensive index calculation unit 51 Color space conversion unit 52 Absolute value calculation unit 53 Color difference calculation unit 54 Variation calculation unit 55 Spot calculation unit F Subject's face R ₁ Analysis area S Skin tissue (skin) )

Claims (31)

Get the first image of the surface of the skin,
From the first image, a plurality of skin indexes showing predetermined properties of the skin surface are calculated.
The number density of the capillaries is calculated based on the plurality of skin indexes calculated by using the estimation formula obtained in advance for estimating the number density of the capillaries existing in the papillary process inside the skin. How to estimate the internal structure of the skin. Further, a skin comprehensive index is calculated from the plurality of skin indexes, and the skin comprehensive index is calculated.
The method for estimating the internal structure of the skin according to claim 1, wherein the number density of the capillaries is calculated from the comprehensive skin index using the estimation formula. The estimation formula is
The number densities of the second images of a plurality of different skins and the capillaries corresponding to the second images are obtained in advance.
A plurality of the skin indexes are calculated from each of the plurality of second images, and a plurality of the skin comprehensive indexes are calculated based on the calculated skin indexes.
The internal structure of the skin according to claim 2, which is obtained by regression analysis using the plurality of skin comprehensive indexes and the calculated number densities of the capillaries corresponding to each of the plurality of skin comprehensive indexes. Estimating method. The first image is an uneven image including information on the unevenness of the surface of the skin.
The predetermined property is wrinkles,
The skin index is a wrinkle index showing the characteristics of the wrinkles.
The uneven image of the surface of the skin is acquired, and the unevenness image is acquired.
In the uneven image, uneven portions on the surface of the skin having a certain depth or a certain area or more are extracted as the wrinkles, and a plurality of the wrinkle indexes are calculated based on the extracted wrinkles.
The method for estimating the internal structure of the skin according to claim 1, wherein the number density of the capillaries is calculated based on the calculated wrinkle index using the estimation formula. The method for estimating the internal structure of the skin according to claim 4, wherein the wrinkle index is calculated based on at least two of the number of wrinkles, the length of wrinkles, the area of wrinkles, the depth of wrinkles, and the angle of wrinkles. .. The wrinkle index calculated based on the length of the wrinkles is calculated by calculating the long axis when the peripheral edge of each wrinkle is approximately elliptical for a plurality of wrinkles extracted from the uneven image as the length of each wrinkle. The method for estimating the internal structure of the skin according to claim 5, which is the calculated average value or maximum value of the lengths of the plurality of wrinkles. The wrinkle index calculated based on the area of the wrinkles is calculated by calculating the number of pixels in the wrinkle region occupied by each wrinkle as the area of each wrinkle for a plurality of wrinkles extracted from the uneven image. The method for estimating the internal structure of the skin according to claim 5 or 6, which is the average value or the maximum value of the wrinkle area. The wrinkle index calculated based on the depth of the wrinkles is the average depth of the pixel values in the wrinkle region occupied by each wrinkle for each of the plurality of wrinkles extracted from the uneven image. The method for estimating the internal structure of the skin according to any one of claims 5 to 7, which is an average value obtained by averaging the average depth of each of the calculated wrinkles with the plurality of wrinkles. The wrinkle index calculated based on the angle of the wrinkles is an angle formed by the long axis of the plurality of wrinkles extracted from the unevenness image in the horizontal direction of the unevenness image when the peripheral edge of each wrinkle is approximately elliptical. The method for estimating the internal structure of the skin according to any one of claims 5 to 8, which is calculated as the wrinkle angle and is the average value of the calculated wrinkle angles or the standard deviation value. The estimation formula estimates the number density of capillaries existing in the papillary process inside the skin corresponding to the wrinkle comprehensive index based on the wrinkle comprehensive index calculated from the plurality of wrinkle indexes. ,
From the wrinkle comprehensive index obtained based on the plurality of wrinkle indexes calculated from the uneven image of the surface of the skin, the number density of capillaries existing in the papillary process inside the skin is used using the estimation formula. The method for estimating the internal structure of the skin according to any one of claims 4 to 9. The estimation formula is
Acquires unevenness images that include unevenness information on multiple different skin surfaces,
In each of the unevenness images including the unevenness information on the surface of the plurality of different skins, the uneven portion having a certain depth or a certain area or more is calculated as a wrinkle.
Based on the extracted wrinkles, a plurality of wrinkle indexes indicating the characteristics of the wrinkles are calculated in each of the images including the unevenness information of the surface of the plurality of different skins.
Based on the calculated plurality of wrinkle indexes, the wrinkle comprehensive index is calculated for each of the images including the unevenness information of the plurality of different skin surfaces.
Obtain the number density of capillaries present in the papillary processes inside the plurality of different skins corresponding to each of the plurality of the wrinkle comprehensive indexes.
The item according to any one of claims 4 to 10, which is obtained by regression analysis using the plurality of wrinkle comprehensive indexes and the number densities of the capillaries corresponding to each of the plurality of wrinkle comprehensive indexes. The method for estimating the internal structure of the skin described. The wrinkle comprehensive index is
The method for estimating the internal structure of the skin according to claim 10 or 11 , which is calculated by the linear sum of the plurality of wrinkle indexes. The method for estimating the internal structure of the skin according to any one of claims 4 to 12, wherein the uneven image is obtained by imaging the skin surface from a predetermined distance. The method for estimating the internal structure of the skin according to any one of claims 4 to 13, wherein the uneven image is obtained by photographing a replica of the skin surface. The method for estimating the internal structure of the skin according to any one of claims 4 to 14, wherein the uneven image is acquired by using a confocal microscope, a light cutting method, or a Gray code pattern light projection method. The first image is an image of the surface of the skin.
The predetermined property is dullness,
The skin index is a plurality of dullness indexes showing the characteristics of the dullness.
An image of the surface of the skin is acquired and
A plurality of the dullness indexes are calculated from the image of the surface of the skin.
Using the estimation formula, the number density of the capillaries is calculated based on the calculated dullness index.
The method for estimating the internal structure of the skin according to claim 1. An analysis area is set in the image, and at least one of brightness, saturation, and hue angle is calculated from the analysis area.
The method for estimating the internal structure of skin according to claim 16, wherein the dullness index is calculated based on at least one of the calculated brightness, saturation, and hue angle. The method for estimating the internal structure of skin according to claim 17, wherein the dullness index includes at least one absolute value of the brightness, the saturation, and the hue angle. The dullness index is
Further, at least one of the brightness, the saturation, and the hue angle is calculated from the skin region excluding the analysis region.
The method for estimating the internal structure of skin according to claim 17 or 18, which includes a difference between the analysis region and the skin region, which is acquired based on at least one of the brightness, the saturation, and the hue angle. The skin according to any one of claims 17 to 19, wherein the dullness index includes a standard deviation obtained based on at least one of the brightness, the saturation, and the hue angle calculated from the analysis region. How to estimate the internal structure. The dullness index is
Any one of claims 17 to 20, which includes at least one of the number of stains and the area ratio of stains obtained based on at least one of the brightness, the saturation, and the hue angle calculated from the analysis region. The method for estimating the internal structure of the skin described in the section. The method for estimating the internal structure of the skin according to any one of claims 17 to 21, wherein the analysis region is an eye region. The estimation formula estimates the number density of the capillaries corresponding to the dullness comprehensive index based on the dullness comprehensive index calculated from the plurality of dullness indexes.
Any one of claims 17 to 22 for calculating the number density of the capillaries using the estimation formula from the dullness comprehensive index obtained based on the plurality of dullness indexes calculated from the image of the surface of the skin. The method for estimating the internal structure of the skin described in the section. The estimation formula is
Acquire images containing the dullness of multiple different skins
A plurality of the dullness indexes are calculated from the images including the dullness of the plurality of different skins.
Based on the calculated dullness index, the dullness comprehensive index is calculated for each of the images including the dullness information of the plurality of different skin surfaces.
Obtain the number density of capillaries present in the papillary processes inside the plurality of different skins corresponding to each of the plurality of the dullness comprehensive indicators.
The estimation of the internal structure of the skin according to claim 23, which is obtained by regression analysis using the plurality of dullness comprehensive indexes and the number densities of the capillaries corresponding to each of the plurality of dullness comprehensive indexes. Method. The dullness comprehensive index is
The method for estimating the internal structure of the skin according to claim 23 or 24, which is calculated by the linear sum of the plurality of dullness indexes. The steps to get the first image of the surface of the skin,
From the first image, a step of calculating a plurality of skin indexes showing predetermined properties of the skin surface, and
The number density of the capillaries is calculated based on the plurality of skin indexes calculated by using the estimation formula obtained in advance for estimating the number density of the capillaries existing in the papillary process inside the skin. Steps to do and
A program that estimates the internal structure of the skin to allow a computer to run. The first image is an uneven image including information on the unevenness of the surface of the skin.
The predetermined property is wrinkles,
The plurality of skin indexes are a plurality of wrinkle indexes indicating the characteristics of the wrinkles.
The step of acquiring the uneven image of the surface of the skin and
In the uneven image, a step of extracting uneven portions on the surface of the skin having a certain depth or a certain area or more as the wrinkles and calculating the plurality of wrinkle indexes based on the extracted wrinkles.
A step of calculating the number density of the capillaries based on the calculated wrinkle index using the estimation formula, and
26. The program for estimating the internal structure of the skin according to claim 26. The first image is an image of the surface of the skin.
The prescribed symptom is dullness
The plurality of skin indexes are a plurality of dullness indexes indicating the characteristics of the dullness.
The step of acquiring the image of the surface of the skin and
A step of calculating a plurality of the dullness indexes from the image of the surface of the skin, and
A step of calculating the number density of the capillaries based on the plurality of dullness indexes calculated by using the estimation formula, and
26. The program for estimating the internal structure of the skin according to claim 26. An image acquisition unit that acquires the first image of the surface of the skin,
A skin index calculation unit that acquires a plurality of skin indexes indicating predetermined properties of the skin surface from the first image, and
Based on the plurality of skin indexes calculated in advance using an estimation formula for estimating the number density of capillaries existing in the papillary process inside the skin, the number density of the capillaries is determined. Capillary number density estimation unit to be calculated and
A device for estimating the internal structure of the skin. The image acquisition unit is a concavo-convex image acquisition unit that acquires a concavo-convex image including information on concavities and convexities on the surface of the skin as the first image.
The skin index calculation unit is a wrinkle index calculation unit that calculates a plurality of wrinkle indexes showing the characteristics of wrinkles as the predetermined properties from the uneven image.
The capillary number density estimation unit calculates the number density of the capillaries based on the plurality of wrinkle indexes calculated by using the estimation formula.
The wrinkle index calculation unit extracts uneven portions on the surface of the skin having a certain depth or a certain area or more as the wrinkles in the uneven image, and based on the extracted wrinkles, the plurality of wrinkle indexes. 29. The device for estimating the internal structure of the skin according to claim 29. The image acquisition unit acquires an image of the surface of the skin as the first image.
The skin index calculation unit is a dullness index calculation unit that calculates a plurality of dullness indexes showing the characteristics of dullness as the predetermined properties from the image of the skin surface.
The internal structure of the skin according to claim 29, wherein the capillary number density estimation unit calculates the number density of the capillaries based on the plurality of dullness indexes calculated by using the estimation formula. Estimator.