WO2009142069A1

WO2009142069A1 - Method for automatically judging skin texture and/or crease

Info

Publication number: WO2009142069A1
Application number: PCT/JP2009/056892
Authority: WO
Inventors: 和広山崎; 平井　義和; 小林　宏
Original assignee: ポーラ化成工業株式会社; 学校法人東京理科大学
Priority date: 2008-05-23
Filing date: 2009-04-02
Publication date: 2009-11-26
Also published as: CN102036607A; TWI454736B; RU2010152571A; HK1153920A1; RU2470576C2; CN102036607B; JP5263991B2; TW200949291A; JPWO2009142069A1

Abstract

A technique for judging the skin texture of a person and/or the crease thereof quickly with high accuracy and a method for selecting a skin external medicine on the basis the result of the judgment are provided. A skin texture and/or crease judging method comprises a step of carrying out image processing of a captured skin image including cross binarization and/or short-line segment matching and thereby obtaining a physical quantity of the skin and a step of substituting the obtained physical quantity into a prepared prediction formula to obtain an evaluation value and thereby judging the difference between the obtained evaluation value and the evaluation value of the skin texture and/or crease. The skin texture and/or crease judging method realizes the technique and the selecting method.

Description

Automatic identification of skin texture and / or wrinkles

The present invention relates to a technique for discriminating the state of the skin, and more particularly, to a technique for discriminating skin texture and / or wrinkles using the physical quantity of the skin as an index.

The important thing when using cosmetics is to accurately select and use cosmetics suitable for the user, and at the same time confirm the effects of the use. If cosmetics that are not suitable for the user's skin are used, not only the effect is not obtained, but also undesirable events may appear. For this reason, the mistake which selects the cosmetics which are not suitable for a user's skin must be avoided at all. Complaints that “cosmetics do not fit the skin” are relatively common complaints, but this phenomenon is also not suitable for the skin. In other words, cosmetics that are inappropriate for the skin It is often due to an error that has been selected. Evaluation of skin texture and wrinkles is an important factor for selecting cosmetics suitable for the skin and confirming the effects of use.

From this point of view, various techniques for selecting cosmetics have been developed. For example, a technique for extracting and analyzing a skin groove pattern by illumination of a skin replica (see Patent Document 1), a technique for analyzing an image directly from the skin surface (see Patent Document 2), and a replica photographed image , Technology for performing image processing such as gradation processing (see Patent Document 3), technology for measuring the depth of wrinkles from a replica using electromagnetic waves (see Patent Document 4), and wrinkles from a skin replica agent using a standard scale Technology for measuring depth (see Patent Literature 5), technology for thinning a single color pixel image or processed pixel image in the skin state, and measuring texture using the peak width interval of the fine lines as an index (Patent Literature 6) Reference) and the like. However, a major problem in the techniques for selecting cosmetics that have been known so far is whether the result of discrimination of texture and / or wrinkles and the evaluation value of apparent texture and / or wrinkles sufficiently match. Furthermore, it is the point that how quickly evaluation can be performed at the site of sales and counseling. In other words, the objectivity and high speed of the evaluation are compatible. In other words, the evaluation value of the texture and / or wrinkles of the appearance is determined by the subjective observation of the replica etc. with the naked eye under a microscope, even if there is a determination criterion. In addition to maintaining objectivity, The length of time required for the determination was a serious problem. For this reason, a discrimination technique having objectivity and high speed and capable of obtaining texture and / or wrinkle evaluation has been desired.

Under such circumstances, filter processing, binarization processing, image signal processing, or matching processing, which are image analysis techniques that have been generally performed in the past, are characteristic of the shape of the skin crevice and hill. Sufficient information cannot be extracted and quantified. Therefore, the present inventors have found a quantification technique (see Patent Document 7) using a cross-binarization process and a short straight line matching, which is a completely new technique for extracting target information from the target image. It was.

JP 60-053121 A JP-A-64-059145 Japanese Patent Laid-Open No. 02-046833 Japanese Patent Laid-Open No. 08-145635 JP 2000-342556 A JP 2006-061170 A JP 2008-061892 A

The present invention has been made under such circumstances, and it is an object of the present invention to provide a technique for distinguishing between skin texture and / or wrinkles that can quickly and accurately distinguish skin texture and / or wrinkles. And It is another object of the present invention to provide a method for selecting an external preparation for skin based on the result of the discrimination.

The inventors of the present invention have intensively researched a skin texture and / or wrinkle discrimination method capable of quickly and accurately discriminating skin texture and / or wrinkles. Skin texture and / or wrinkles are identified with high accuracy and quickly by substituting the physical quantity of the skin obtained by performing the skin treatment and / or the short straight line matching processing into an estimation formula prepared in advance. We found out what we could do and came to complete the invention. That is, this invention relates to the technique shown below.
(1) Performing a cross binarization process and / or a short straight line matching process on the skin image to obtain the physical quantity of the skin, and substituting the physical quantity of the skin obtained in the process into a prediction formula prepared in advance. A method for distinguishing skin texture and / or wrinkles, comprising a step of distinguishing the obtained evaluation value from an evaluation value of skin texture and / or wrinkles.
(2) Means for inputting a prediction formula prepared in advance, means for acquiring a skin image, means for calculating a physical quantity of skin from the acquired skin image, a prediction formula prepared in advance and the physical quantity of the calculated skin An apparatus for distinguishing skin texture and / or wrinkles, comprising: means for calculating an evaluation value of skin texture and / or wrinkles; and means for displaying the calculated evaluation value.
(3) Let the computer function as means for calculating a physical quantity from the acquired skin image, a prediction formula prepared in advance, and means for calculating an evaluation value of skin texture and / or wrinkles from the physical quantity of the skin Skin discrimination program.
(4) The step of discriminating skin texture and / or wrinkles using the discrimination method according to (1) or the discrimination device according to (2), and the skin texture and / Or, based on the evaluation value of the wrinkle, in the case of the discrimination result that the skin texture and / or wrinkle state of the subject is not good, the topical skin containing ingredients for improving the texture state and preventing the disorder of the texture state A method for selecting an external preparation for skin, comprising a step of selecting an external preparation and, in the case of a discrimination result that the skin texture and / or wrinkle state of the subject is good, selecting an external preparation for skin containing only a moisturizing component.

According to the present invention, it is possible to provide a technique for distinguishing skin texture and / or wrinkles that can quickly and accurately distinguish skin texture and / or wrinkles. Moreover, the technique can be applied and the skin external preparation suitable for a user can be provided.

It is a figure which shows the brightness | luminance image of a replica (drawing substitute photograph). It is a figure which shows the cross binarization processing method (drawing substitute photograph). It is a figure which shows the image of the replica by a cross binarization process (drawing substitute photograph). It is a figure which shows a short straight line matching processing method (drawing substitute photograph). It is a figure which shows the image of the replica by a short straight line matching process (drawing substitute photograph). It is a figure which shows the image (left: evaluation value 1, right: evaluation value 5) of the reference | standard photograph of a texture (drawing substitute photograph). It is a figure which shows the image (left: evaluation value 1, right: evaluation value 3) of the wrinkle reference | standard photograph (drawing substitute photograph). It is a figure which shows the area of the skin groove which is a physical quantity after a cross binarization process and a short straight line matching process (drawing substitute photograph). It is a figure which shows the skin groove average thickness which is a physical quantity after a cross binarization process and a short straight line matching process (drawing substitute photograph). It is a figure which shows the structural example of a discrimination apparatus. It is a flowchart which concerns on the process of a discrimination apparatus.

In the discrimination method of the present invention, it is possible to discriminate the texture of the skin, the wrinkle of the skin, or both from the skin image.

<Acquisition of skin image used in the present invention>
In the present invention, a skin image is used. The method for acquiring the skin image may be a method for directly capturing the skin and obtaining the skin image, or a method for obtaining the skin image through a replica specimen collected from the skin. As a method for acquiring an image, for example, a digital video camera can be used through a stereomicroscope, or a commercially available digital microscope can be used. Examples of such a digital microscope include a cosmetic microscope manufactured by Moritex Co., Ltd. and a digital microscope manufactured by Keyence Co., Ltd.

In the present invention, it is preferable to obtain a skin image through a replica specimen collected from the skin. By obtaining a skin image through a replica specimen collected from the skin, it is possible to exclude only color information on the surface of the skin and obtain only morphological information, thereby preventing noise such as spots. In addition, unevenness on the skin surface that is not necessary for analysis (not unevenness on the skin groove or hill level) is canceled by replica collection, so that the analysis becomes easy. A method for obtaining a skin image through a replica specimen will be described below.

A microscope lens is arranged at a 90 ° position with respect to the replica specimen, light is irradiated to the replica specimen at an appropriate angle, and a shadow image of the surface irregularities of the replica specimen caused by incident light is captured as an image through the microscope. be able to. Here, the replica specimen is a solvent-softening transparent plastic plate that is coated with a softening solvent and softened, and then the softened part is pressed onto the skin, and the unevenness on the skin is applied to the softened part. It is a technique for indirectly observing the unevenness on the skin by transferring and observing the unevenness, and the “Kawai method” is known as a representative technique. This technique has been widely used for a long time in the field of cosmetic science because the unevenness of the skin is faithfully transferred and the unevenness of the unevenness is excellent. There are already commercially available kits for producing such replica specimens, which can be used. Such a replica specimen is preferably collected from the cheek, or from the corner of the eye corner to the lower portion of the corner of the eye 1.5 cm × 1.5 cm. In a general method, such a replica specimen is generally irradiated with light from a lower surface perpendicular to the replica surface, and transmitted light is observed. That is, the unevenness is observed as an image by utilizing the fact that the irradiated light is scattered by the transferred unevenness and the amount of transmitted light is reduced. In the discrimination method of the present invention, the replica is preferably observed as follows. The uneven surface of the replica is directed in the image-capturing direction, and light is irradiated to this surface at an angle of 10 to 40 degrees, more preferably 20 to 30 degrees. Capture). This is because by adopting such a form, the unevenness transferred to the replica surface appears more clearly as a light intensity difference. Table 1 shows the evaluation of clarity when the same sample is observed with different incident angles. The evaluation criteria are ○: clear, Δ: slightly unclear, ×: unclear.

<Cross-binarization as image processing>
In the present invention, image processing including cross binarization processing and / or short straight line matching processing is performed on the obtained skin image. Such image processing is described in Japanese Patent Laid-Open No. 2008-061892 (Patent Document 7), and will be described below.

As the most basic image processing method, there is a binarization processing method that separates a background and an object from an image and extracts the object as a shape. If there is sufficient contrast between the target and the background, the binarization process is easy. However, in practice, since there is a subtle change in shading mainly at the boundary between the object and the background, it is difficult to set a threshold value for binarization processing for the purpose of highly accurate shape extraction. Also, even when the background gray level varies due to uneven illumination, accurate shape extraction is difficult with a threshold fixed to the entire screen. In such a case, dynamic threshold processing (variable threshold processing) that changes the threshold value for each pixel instead of a fixed threshold value is preferable, and this cross binarization processing method belongs to the dynamic threshold processing method. The processing area of the dynamic threshold processing method is generally rectangular, but this cross binarization processing method has a feature of a cross shape suitable for extracting the shape of the skin groove (see FIG. 2). By using this cross-binarization processing method, shadows formed by the convex parts of the skin groove can be detected without being affected by uneven illumination of the replica, and the entire screen from thick and clear skin grooves to fine skin grooves Therefore, a highly accurate cross-binarized image (see FIG. 3) can be obtained without unevenness.
The cross binarization process can be performed using an epidermis tissue quantification apparatus described in Japanese Patent Application Laid-Open No. 2008-061892.

<Short straight line matching as image processing>
The short straight line matching method is a method for calculating a physical quantity of an object shape in a binarized image. While the conventional method measures the number of target pixels in units of one pixel of the binarized image and calculates physical quantities such as area, length, and center of gravity, this short straight line matching method is composed of a plurality of pixels. The physical quantity is calculated using a short straight line (a length of several pixels to several tens of pixels and a width of one pixel) as a unit. Specifically, if the end point of the target area is the start point of the short line and the end point of the short line is within the target area, the next short line is connected with the end point as a new start point. If the end point of the short straight line is outside the target area, the connection is terminated. This operation is repeated until the target area is covered with a short straight line. Thereafter, the number of short straight lines fitted to the target region, the angle, and the like are measured, and the feature amount of the target object is calculated (see FIG. 4). By this method, it is possible to obtain a short straight line matching image of a feature of a skin groove having directionality that is continuous long and narrow (see FIG. 5).
The short straight line matching process can be performed using an epidermis tissue quantification apparatus described in Japanese Patent Application Laid-Open No. 2008-061892.

In the above image processing, only one of the image processing may be performed, but the physical quantity can be calculated more accurately by performing both image processing. In addition, other image processing such as luminance conversion processing, binarization processing, filter processing, general image processing (area, perimeter, aspect ratio, center of gravity, needle ratio, enlargement, inversion) as necessary. You may go.

<Calculation of physical quantity of texture and / or wrinkles>
In the present invention, the physical quantity of the skin image can be obtained by performing image processing including the cross binarization processing and / or the short straight line matching processing. The obtained physical quantity is a physical quantity obtained by quantifying the features such as the skin groove and skin. Such physical quantities include, for example, physical quantities such as skin groove area, skin groove average thickness, skin groove thickness variation, skin groove spacing, skin groove parallelism, skin groove direction, skin groove density, etc. First, the maximum number of short straight lines for each angle at 95 ° or more, the maximum number of short straight lines for each angle of 10 ° to 90 °, the maximum number of short straight lines for each thickness, More detailed physical quantities, such as the thickness of the number of short straight lines for each thickness, the total value of the short straight line frequency data, the total value of the thickness of the number of short straight lines for each thickness, etc. In the present invention, a physical quantity that is considered to be closely related to texture and wrinkles is calculated from these physical quantities. Specifically: skin groove area = area occupied by skin groove or total number of matching short lines in the image range to be processed; skin groove average thickness = (total sum of skin groove thickness for each matching start point / Total number of starting points); variation in skin groove thickness = standard deviation or variance calculated from histogram of thickness and number of skin grooves; average interval of skin grooves = 1 / (area of skin groove / average skin groove) (Thickness); parallelism of skin groove = concentration or dispersion of peaks calculated from histogram of skin groove angle and number; direction and density of skin groove = number of short straight lines at angle θ (histogram height) / skin It can be defined as the total length of the groove. Other physical quantities can be appropriately calculated from the above calculation formulas. There are a large number of physical quantities obtained in the present invention as described above. From among them, a preferable physical quantity is selected in order to calculate a prediction formula described later. Such calculation of the physical quantity can be processed by a program on a computer including image processing including the cross binarization processing and / or the short straight line matching processing.

<Prediction formula>
In order to discriminate the texture and / or wrinkles of the skin, a prediction formula indicating the relationship between the physical quantity of the skin and the visual evaluation values of the texture and / or wrinkles of the skin is obtained in advance. The prediction formula can be created by the following method, for example.

目視 Visual evaluation of texture and / or wrinkles by the evaluator is performed on skin replicas (hereinafter referred to as samples) with sufficient consideration of skin condition and age. On the other hand, for the sample, the physical quantity of the skin is calculated by the method described above. The number of samples is preferably 100 or more, more preferably 500 or more. The visual evaluation of texture and / or wrinkles is suitable for representing a third party with reference to a 3 to 10-level reference photo to judge whether the texture is good to bad or low to wrinkle to many A plurality of evaluators, preferably 5 or more, are allowed to evaluate the sample and have an evaluation value corresponding to the reference photograph. The evaluator suitable for representing the third party is preferably one who has experience in beauty, aesthetics or skin evaluation research for more than one year and is continuously conducting skin evaluation training. . The average value obtained by excluding the maximum value and the minimum value of the evaluation value of each sample is taken as the visual evaluation value of texture and / or wrinkle of the sample.

The skin texture and the level of wrinkles, which are evaluations of the skin condition, are discussed at the Japan Cosmetic Engineers Association and the International Federation of Cosmetic Engineers (IFSCC), and can be objectively recognized by third parties. As the degree of wrinkles, it is under the shared recognition of most. FIG. 6 and FIG. 7 show an example of a standard photograph of texture (five-level evaluation) and wrinkle (three-level evaluation) standardized based on statistical processing. In this technical field, such a reference photo can be determined to be a reference photo having a certain degree of reliability if the population serving as a basis for creation is about 100, and the population exceeds 1,000. In this case, it can be determined that the reference photo has a considerably high reliability, and it is not necessary to consider the difference between the reference photos. In the sample evaluation in the present invention, a reference photo standardized based on such statistical processing can be used, and the population in the creation of the reference photo is preferably 1,000 or more.

Next, a prediction formula is derived from the visual evaluation value of texture and / or wrinkles thus obtained and the calculated physical quantity of the skin. The prediction formula can be a formula obtained by multivariate analysis of the physical quantity of skin and the visual evaluation value of texture and / or wrinkles as the prediction formula of the present invention. As multivariate analysis, those that can use the relationship between explanatory variables and objective variables are preferable. For example, discriminant analysis, principal component analysis, factor analysis, quantification theory class 1, quantification theory class 2, quantification theory class 3, Regression analysis (MLR, PLS, PCR, logistic), multidimensional scaling method, supervised clustering, neural network, ensemble learning method, etc. can be illustrated, and a prediction formula is created using free software or commercially available one Can do. Of these, multiple regression analysis, discriminant analysis, and quantification theory are particularly preferable. It is preferable to perform multiple regression analysis by performing multiple regression analysis using the texture and / or wrinkle visual evaluation value obtained above as an objective variable, using the physical quantity of skin as an explanatory variable, and using the multiple regression formula as a prediction formula It can be illustrated.

As described above, there are various physical quantities in the physical quantity of the skin used in the calculation of the prediction formula, but from the viewpoint of improving the accuracy of this discrimination method, it is preferable to contain physical quantities related to the skin groove, and the accuracy of the discrimination method is further increased. From the viewpoint of improving, it is more preferable to contain physical quantities relating to 10 or more kinds of skin grooves. The total number of physical quantities used for calculating the prediction formula is preferably 10 or more.

<Difference process>
By substituting the physical quantity of the skin into the prediction formula set in this way and obtaining an evaluation value, it is possible to distinguish between skin texture and / or wrinkles. By assigning the physical quantity of the skin calculated from the obtained image to the prediction formula, a visual evaluation value of skin texture and / or wrinkles can be obtained. The present invention can identify skin texture and / or wrinkles with extremely high accuracy through the above-described steps. Furthermore, the physical quantity and visual evaluation value of a new sample are incorporated into a database, and updated and corrected, so that the accuracy of the prediction formula is further improved and high-precision discrimination is expected.

<Difference device / program>
Another aspect of the present invention is a program for performing the above steps. That is, the computer functions as a means for calculating a physical quantity from the acquired skin image, a prediction formula prepared in advance, and a means for calculating an evaluation value of skin texture and / or wrinkles from the calculated physical quantity of skin. It is a skin discrimination program. The discrimination program of the present invention can be used by installing it on hardware such as a personal computer.
Furthermore, another aspect of the present invention is a discrimination apparatus that performs the above steps. That is, a means for inputting a prediction formula prepared in advance, a means for acquiring a skin image, a means for calculating a physical quantity of the skin from the acquired skin image, a skin from the prediction formula prepared in advance and the calculated physical quantity of the skin An apparatus for distinguishing skin texture and / or wrinkles, comprising: means for calculating an evaluation value of the texture and / or wrinkle of the skin; and means for displaying the calculated evaluation value.

The aspect of the discrimination device will be described with reference to FIG. The identification device of the present invention may be a general-purpose computer such as a personal computer or a dedicated computer for identification. The input unit 1 is an input unit for the prediction formula, and inputs a prediction formula to be used for discrimination in advance. For example, an input device such as a keyboard can be used. The image acquisition unit 2 is a means for acquiring a skin image, and a digital video camera or a commercially available digital microscope can be used. CPU 3 (Central Processing Unit) is a means for calculating the physical quantity of skin from the acquired skin image, and a means for calculating an evaluation value of skin texture and / or wrinkles from the prediction formula prepared in advance and the calculated physical quantity of skin. is there. By installing the above discrimination program, it functions as such means. A RAM 4 (Random Access Memory) is a storage means for storing temporary data. The display unit 5 is a means for outputting the calculated evaluation value, and can be, for example, a display device such as a liquid crystal display or an output device such as a printer.

The processing flow of the discrimination device will be described with reference to FIG.
First, a skin image is acquired from an image acquisition unit such as a digital video camera. As already explained, it can also be taken directly from the skin of the subject or via a replica specimen. The acquired skin image is subjected to image processing such as cross binarization processing and short straight line matching processing in the CPU, and the physical quantity of the skin image is also calculated. The type of the physical quantity of the skin image to be calculated is appropriately set depending on the type of the physical quantity used for calculating the prediction formula input in advance from the input unit. The calculated physical quantity of the skin image is also substituted into a prediction formula input in advance in the CPU, and its evaluation value is calculated. The calculated evaluation value is output from output means such as a liquid crystal display.

In the present invention, it is also possible to discriminate between skin texture and wrinkles at a time by previously inputting a prediction formula for skin texture and a prediction formula for skin wrinkle into the above-described discrimination device. It can also be differentiated separately.

<Selection method of external preparation for skin based on identified skin texture evaluation value>
Based on the evaluation value of the texture discriminated by the discrimination method or the discrimination device, a skin external preparation suitable for the subject having the skin image used can be selected. By using the discrimination method or the discrimination device of the present invention, it is possible to quickly discriminate with high accuracy almost the same as when an expert evaluates the skin, and based on the result, the skin texture state is maintained. Therefore, a skin external preparation useful for prevention or improvement can be selected.

In the selection of external preparations for skin, especially in the selection of cosmetics, if a display of a differential value indicating that the skin texture of the subject is not good is output, in order to improve the texture condition or prevent disturbance of the texture condition By selecting a cosmetic containing these components, a cosmetic suitable for the skin of the subject can be selected. Examples of such components include turnover promoting components, collagen synthesis promoters, stratum corneum detachment promoters, collagen fiber bundle restructuring agents, and the like, and one or more of these can be contained. . Of these, the collagen fiber bundle restructuring agent is most effective in improving texture.

As the above-mentioned turnover promoting component, retinoic acid, phytosteside, phytosterol, sphingosine, steroid and the like can be mentioned. Examples of the collagen synthesis promoter include bakugankon extract. Examples of the stratum corneum elimination promoter include α-hydroxy acids. Examples of the collagen fiber bundle restructuring agent include rosemary extract and ursolic acid derivative.

On the other hand, when the display of the discrimination value that the condition of the texture of the subject's skin is good is output, the condition of the texture can be maintained by selecting the cosmetic containing only the moisturizing component, An appropriate cosmetic amount can be selected. Examples of the moisturizing component include heparin-like substances. These components may be contained so as to exhibit the effects when contained in cosmetics, and preferably 0.01 to 5.0% by mass, respectively. Appropriate cosmetics can be selected by a person skilled in the art appropriately selecting these components based on the displayed texture discrimination values. As an example, selection examples of cosmetic ingredients with respect to the texture discrimination value in five stages are shown below. In addition, naturally these ingredients can be contained not only in cosmetics but also in other skin external preparations.
<Kime discrimination value-cosmetic ingredient>
1 (Good)-Moisturizing component 2-Collagen synthesis promoter, moisturizing component 3-Collagen synthesis promoter, stratum corneum release promoter, moisturizing component 4-Collagen fiber bundle restructuring agent, collagen synthesis promoter, moisturizing component 5 ( Evil)-Collagen fiber bundle restructuring agent, collagen synthesis promoter, stratum corneum detachment promoter, moisturizing ingredient

<Selection method of external preparation for skin based on differentiated skin wrinkle evaluation value>
Based on the evaluation value of wrinkles discriminated by the discrimination method or the discrimination device, it is possible to select a skin external preparation suitable for the subject having the differentiated skin image, as in the case of texture. As an example of cosmetics to be selected based on the evaluated evaluation value of wrinkles, examples of selecting cosmetic ingredients for the three-stage wrinkle discrimination values are shown below.
<Wrinkle discrimination value-cosmetic ingredient>
1 (good)-moisturizing component 2-collagen synthesis promoter, stratum corneum detachment promoter, moisturizing component 3 (bad)-collagen fiber bundle restructuring agent, collagen synthesis promoter, horny layer detachment promoter, moisturizing component

In the selection of the above cosmetics, it can be used in combination with other indicators such as not only texture or wrinkle discrimination values but also various skin characteristic values, observation results of skin surface morphology, and palatability. Combinations with such other indicators also belong to the technical category of the present invention. As such other indicators, specifically, for example, skin characteristic values such as transepidermal water transpiration (TEWL) and conductance, or characteristic values of stratum corneum cells obtained by tape stripping (for example, cell Area, cell volume, cell area dispersion, cell flatness, cell arrangement regularity, stratum corneum detachment, presence of nucleated cells), water retention function of skin estimated by the characteristic values, sebum secretion Examples are the amount, age of skin, melanin production ability, skin color, skin properties and skin quality. Among these, a water retention function deeply related to the textured state and the wrinkled state is particularly preferable.

Examples of the present invention will be described below, but the scope of the present invention is not limited thereby.

<Process for visual evaluation of texture and wrinkles>
Replica specimens were collected from the center of the cheeks of 30 women in their 10s and 50s, and images were stored as digital data from the replica specimens using a Moritex cosmetic microscope. Using a general-purpose personal computer incorporating software for performing the image processing described above, noise processing is performed on the image to convert it to a luminance image, cross binarization processing and short straight line matching processing are performed, The physical quantity for was calculated. As physical quantities, the skin groove area (see FIG. 8), skin groove average thickness (see FIG. 9), skin groove thickness variation, skin groove spacing, skin groove parallelism, skin groove direction, skin groove density 17 physical quantities were calculated. As can be seen from FIG. 8 and FIG. 9, these physical quantities clearly show the features of the unevenness of the skin grooves and the hills, and it can be seen that they are indexes that are very easy to evaluate visually.

<Automatic identification of texture visual evaluation>
Using a reference photograph (see Fig. 6: created by the inventors based on a population of 1000) of a five-stage evaluation of a replica specimen collected from the center of a woman's cheek, three specialists Digital image of skin texture evaluation value 1 to 5 (1: good to 5: bad) from digital image of skin replica evaluated by skin evaluator and 15,000 pieces of visual evaluation value data thereof A total of 1000 sheets were selected for each evaluation value of 200 sheets. The specialized skin evaluator is a person who has experience in beauty, aesthetics, or skin evaluation research for more than one year, and intermittently performs skin evaluation training. The physical quantity was calculated using the method shown in Example 1 for these 1000 sheets. Next, it is divided into two groups, group A and group B, each having 500 sheets randomly so that the visual evaluation values of 1 to 5 are 100 sheets each, and the purpose is to visually evaluate the texture of 500 sheets of group A Multiple regression analysis (manufactured by SPS Co., Ltd.) was performed using 17 physical quantities as explanatory variables, and a multiple regression equation as a prediction formula was calculated (multiple correlation coefficient = 0.909). For the remaining 500 images of group B, the previously calculated physical quantity was substituted into the explanatory variable of this multiple regression equation, and the visual evaluation value (automatic discrimination value) of the texture was discriminated. A list of physical quantities used is shown in Table 2, and the results are shown in Table 3.

Table 3 shows an aggregation table of the texture visual evaluation value (automatic discrimination value) obtained by the present invention and the texture visual evaluation value. The Spearman correlation coefficient is 0.887, the perfect match between the two evaluation values is 62%, and 98% when one-step deviation is allowed, indicating that the texture discrimination method of the present invention has a sufficiently satisfactory accuracy.

Visual evaluation value (automatic discrimination) of texture of group A using the multiple regression equation (multiple correlation coefficient = 0.935) prepared for the image of group B by exchanging group A and group B in Example 2 Value). The results are shown in Table 4.

Table 4 shows a tabulation table of texture visual evaluation values (automatic discrimination values) obtained by the present invention and texture visual evaluation values. The Spearman correlation coefficient is 0.861, the perfect match between the two evaluation values is 53%, and 97% when one-step deviation is allowed. From these results, it can be seen that automatic discrimination with high accuracy can always be performed on unknown data.

<Automatic identification of visual evaluation of wrinkles>
In Example 2, a digital image having skin wrinkle evaluation values 1 to 3 was selected in a total of 600 images, each having 200 evaluation values, and the same procedure as in Example 2 was performed. The multiple correlation coefficient of the multiple regression equation obtained by the multiple regression analysis is 0.912, and the Spearman correlation coefficient between the visual evaluation value of wrinkles (automatic discrimination value) and the visual evaluation value of wrinkles is 0.705, The perfect match between the two evaluation values is 65% and 100% when one step deviation is allowed, and it can be seen that the wrinkle discrimination method of the present invention has a sufficiently satisfactory accuracy.

In Example 4, as a result of exchanging data of Group A and Group B and performing automatic discrimination in the same manner, the multiple correlation coefficient of the obtained multiple regression equation = 0.820), visual evaluation value of wrinkles (automatic discrimination) Value) and the visual evaluation value of wrinkles have a Spearman correlation coefficient of 0.880, the perfect match between the two evaluation values is 84%, and 100% when one-step deviation is allowed. From these results, it can be seen that even in the evaluation of wrinkles, it is possible to always perform automatic discrimination with high accuracy for unknown data.

<Automatic identification of texture visual evaluation>
In Example 2, the multiple regression analysis was replaced with a neural network (manufactured by NeuralWare) and used for supervised learning for the group A, and learning was performed with physical quantities using the textured visual evaluation value as a response variable to obtain a prediction formula. . The physical quantity of the skin was substituted into the obtained prediction formula, and the visual evaluation value (automatic discrimination value) of texture of Group B was discriminated. The results are shown in Table 5.

Table 5 shows a tabulation table of the texture visual evaluation value (automatic discrimination value) obtained in the present invention and the texture visual evaluation value. The Spearman correlation coefficient is 0.871, the perfect match between the two evaluation values is 62%, and 99% when one-step deviation is allowed. From these results, it can be seen that even when a prediction formula is created using a multivariate analysis method other than multiple regression analysis, automatic discrimination with high accuracy can be performed.

In Example 2, using the three physical quantities shown in Table 6 as explanatory variables, multiple regression analysis (manufactured by SPS Corporation) was performed to calculate a multiple regression equation (multiple correlation coefficient = 0.880). ). The Spearman correlation coefficient was 0.831, the perfect match rate between the two evaluation values was 47%, and 95% when one-stage deviation was allowed. Although the discrimination can be performed with high accuracy even if the number of physical quantities is small, it can be seen that the accuracy increases as the number of physical quantities increases.

Tables 7 and 8 show the agreement rate of evaluation and the required time (seconds) of evaluation per sample for Examples and Comparative Examples for distinguishing texture and wrinkles. That is, texture evaluation (comparative example 1) and wrinkle evaluation (comparative example 2) by three specialist skin evaluators (trainers), which are the standards for visual evaluation, and non-trainers (standards of FIGS. 6 and 7). It is texture evaluation (Comparative Example 3) and wrinkle evaluation (Comparative Example 4) by explaining and using a photograph. Further, in the first and second embodiments, the binarization process and the thin line process (without using the cross binarization and the short straight line matching process) (connectivity is added to the connected figure in the binarized image obtained by sampling). To the line figure so as not to be lost), and is a statistical physical quantity generally obtained by the process, which is the total area, standard deviation, total number, unit area, and thin line peak interval, Identification was performed in the same manner as in Example 2 using values, standard deviation, standard error, coefficient of variation, and the like. The visual evaluation values (automatic discrimination values) of the texture and wrinkles were referred to as Comparative Examples 5 and 6, respectively.

<Use test by cosmetic selection method based on automatic texture discrimination value>
A long-term use test of cosmetics was conducted on the effectiveness of the cosmetics selection method based on the texture discrimination value for female subjects.
First, five types of cosmetics (cosmetics 1 to 5) corresponding to the textured condition of the skin were prepared according to the usual cosmetic preparation method based on the following formulation.

(Cosmetics 1: Cosmetics for automatic identification value 1)
Component Content Glycerin 5% by mass
1,3-butanediol 5% by mass
Soy protein 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 5% by mass
Methylparaben 0.1% by mass
Water remaining

(Cosmetics 2: Cosmetic for automatic texture identification value 2) Content Glycerin 5% by mass
1,3-butanediol 5% by mass
Bakugakon Extract 0.1% by mass
Soy protein 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 5% by mass
Methylparaben 0.1% by mass
Water remaining

(Cosmetics 3: Cosmetic for automatic texture discrimination value 3) Content Glycerin 5% by mass
1,3-butanediol 5% by mass
Bakugakon Extract 0.1% by mass
Sodium lactate 0.1% by mass
Soy protein 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 5% by mass
Methylparaben 0.1% by mass
Water remaining

(Cosmetics 4: Cosmetic for automatic texture identification value 4) Content Glycerin 6% by mass
1,3-butanediol 5% by mass
Rosemary extract 0.1% by mass
Bakugakon Extract 0.1% by mass
Soy protein 0.1% by mass
Stearyl ursolate 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 10% by mass
Methylparaben 0.1% by mass
Water remaining

(Cosmetic 5: Cosmetic for automatic texture discrimination value 5) Content Glycerol 7% by mass
1,3-butanediol 5% by mass
Rosemary extract 0.1% by mass
Bakugakon Extract 0.1% by mass
Sodium lactate 0.1% by mass
Soy protein 0.1% by mass
Stearyl ursolate 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 15 mass%
Methylparaben 0.1% by mass
Water remaining

Next, 80 healthy female panelists (aged 32 to 57) randomly selected were divided into two groups, A and B, so that there was no difference in age distribution. In Group A, a texture discrimination value is automatically calculated from the cheek replica specimen using the multiple regression equation obtained in Example 2, and cosmetics 1 to 5 corresponding to the texture automatic discrimination values 1 to 5 are provided. I gave it. The automatic texture discrimination value was calculated in the same manner for the B group, but the cosmetic 5 for the automatic texture discrimination value 5 was handed over ignoring the automatic texture discrimination value. The panelists of both groups A and B were allowed to use the delivered cosmetic for 3 months, and the automatic texture discrimination value was calculated in the same manner after use. The texture improvement value was calculated from “texture improvement value” = “automatic texture discrimination value before use test” − “texture automatic discrimination value after use”, and an average value was obtained for each group. In addition, when the subject complained that the cosmetic used during the test period was “not suitable for the skin”, the use of the cosmetic was interrupted and excluded from the evaluation. The results are shown in Table 9. From Table 9, when the cosmetics selection method of this invention is employ | adopted, it turns out that the person who does not match skin does not exist, and has a texture improvement effect.

<Use test by cosmetic selection method based on automatic wrinkle discrimination value>
A long-term use test of cosmetics was conducted on the effectiveness of the cosmetic selection method based on wrinkle discrimination values for female subjects.
First, based on the formulation shown below, three types of cosmetics (cosmetics 1 to 3) corresponding to the wrinkle state of the skin were prepared according to the usual cosmetic preparation method.

(Cosmetics 1: Cosmetics for automatic wrinkle discrimination value 1)
Component Content Glycerin 5% by mass
1,3-butanediol 5% by mass
Bakugakon Extract 0.1% by mass
Soy protein 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 10% by mass
Methylparaben 0.1% by mass
Water remaining

(Cosmetics 2: Cosmetics for automatic wrinkle discrimination value 2) Content Glycerin 5% by mass
1,3-butanediol 5% by mass
Bakugakon Extract 0.1% by mass
Sodium lactate 0.1% by mass
Soy protein 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 10% by mass
Methylparaben 0.1% by mass
Water remaining

(Cosmetics 3: Cosmetics for automatic wrinkle discrimination value 3) Content Glycerin 7% by mass
1,3-butanediol 5% by mass
Rosemary extract 0.1% by mass
Bakugakon Extract 0.1% by mass
Sodium lactate 0.1% by mass
Soy protein 0.1% by mass
Stearyl ursolate 0.1% by mass
Heparin-like substance 0.1% by mass
Ethanol 15 mass%
Methylparaben 0.1% by mass
Water remaining

Next, 72 healthy female panelists (age 30-58) randomly selected were divided into two groups, A and B, so that there was no difference in age distribution. In group A, wrinkle discrimination values are automatically calculated from cheek replica specimens using the multiple regression equation obtained in Example 4, and cosmetics 1 to 3 corresponding to the wrinkle automatic discrimination values 1 to 3 are provided. I gave it. The wrinkle automatic discrimination value was calculated in the same manner for group B, but the wrinkle automatic discrimination value was ignored and the cosmetic 2 for the wrinkle automatic discrimination value 2 was passed. The panelists of both groups A and B were allowed to use the delivered cosmetics for 6 months, and the wrinkle automatic discrimination value was calculated in the same manner after use. The “wrinkle improvement value” = “wrinkle automatic discrimination value before use test” − “wrinkle automatic discrimination value after use” was calculated, and the wrinkle improvement value was calculated to obtain an average value for each group. The results are shown in Table 10. From Table 10, it can be seen that when the cosmetic selection method of the present invention is adopted, it has a wrinkle improving effect.

The present invention can provide a technique for easily and quickly discriminating skin texture or wrinkles anywhere with ease. As a result, information useful for skin and beauty counseling and cosmetics selection can be provided, for example, at department stores and stores.

Claims

Performing image processing including cross binarization processing and / or short straight line matching processing on a skin image to obtain a physical quantity of the skin;
Substituting the physical quantity of the skin obtained in the step into a prediction formula prepared in advance, and distinguishing the obtained evaluation value from the skin texture and / or wrinkle evaluation value, and skin texture and / or Wrinkle discrimination method.
The skin texture and / or wrinkle according to claim 1, wherein the prediction formula is a formula obtained by multivariate analysis of a physical quantity of skin and a visual evaluation value of texture and / or wrinkles. Identification method.
3. The method for distinguishing skin texture and / or wrinkles according to claim 1 or 2, wherein the physical quantity of the skin includes a physical quantity related to a skin groove.
The method for distinguishing skin texture and / or wrinkles according to any one of claims 1 to 3, wherein the physical quantity of the skin contains physical quantities relating to 10 or more kinds of skin grooves.
The method for distinguishing skin texture and / or wrinkles according to any one of claims 1 to 4, wherein the skin image is a skin image obtained through a skin replica specimen.
6. The skin texture according to claim 5, wherein the skin image is an image obtained by irradiating the replica specimen with light at an angle of 10 to 40 degrees and capturing an image made of the reflected light. And / or wrinkle discrimination method.
Means for inputting a prediction formula prepared in advance;
Means for acquiring a skin image;
Means for calculating a physical quantity of the skin from the acquired skin image;
Means for calculating an evaluation value of skin texture and / or wrinkles from a prediction formula prepared in advance and the calculated physical quantity of the skin;
A skin texture and / or wrinkle discrimination device comprising: means for displaying the calculated evaluation value.
Computer
Means for calculating a physical quantity from the acquired skin image;
A skin discrimination program that functions as a prediction formula prepared in advance and a means for calculating an evaluation value of skin texture and / or wrinkles from the calculated physical quantity of the skin.
A step of discriminating skin texture and / or wrinkles using the discrimination method according to any one of claims 1 to 6 or the discrimination device according to claim 7, and the skin texture discriminated by the discrimination step and In the case of a discrimination result that the texture and / or wrinkle state of the subject's skin is not good based on the evaluation value of the wrinkle, the skin containing a component for improving the texture state and preventing disturbance of the texture state A method for selecting a skin external preparation, comprising a step of selecting an external preparation and, in the case of a discrimination result that the skin texture and / or wrinkle state of the subject is good, selecting a skin external preparation containing only a moisturizing component.
The method for selecting an external preparation for skin according to claim 9, wherein the external preparation for skin is a cosmetic.
The skin according to claim 10, wherein the cosmetic contains one or more selected from the group consisting of a moisturizing component, a collagen synthesis promoter, a keratin detachment promoter, and a collagen fiber bundle restructuring agent. Selection method of external preparation.