Skin age estimation method by tokyo university-method makeup theory

Abstract

To provide a method of correctly calculating a skin age from a face image of a subject.SOLUTION: A skin age calculation method comprises: acquiring a regression equation being a skin age calculation formula, on the basis of the correlation between an actual age and a variable including at least one of (i) a B value of RGB of the forehead and a B value of RGB of the cheek and (ii) a G value of RGB of the cheek, of a face image; and calculating the skin age of a subject on the basis of the regression equation from a variable including at least one of (i) a B value of RGB of the forehead and a B value of RGB of the cheek and (ii) a G value of RGB of the cheek, of the subject.SELECTED DRAWING: Figure 11

Description

[Technical field]

[0001] The present invention relates to a method for calculating skin age, a makeup method for improving skin age with cosmetics, and a method for evaluating the effect of cosmetics in improving skin age. [Background technology]

[0002] Techniques for analyzing skin conditions from facial images of subjects are widely known, and skin pores, age spots, skin color, amount of sebum, porphyrin in keratin plugs, etc. have been used as indicators in the past (Patent Documents 1 and 2). Furthermore, methods for calculating skin age have been reported recently, such as a method for calculating skin age based on intercellular lipids in the stratum corneum of the skin (Patent No. 6128549) and a method for calculating skin age based on the viscoelasticity of the lower dermis (Patent Document 4). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 4717595 [Patent Document 2] Patent No. 5080060 [Patent Document 3] Patent No. 6128549 [Patent Document 4] Patent No. 7254340 Summary of the Invention [Problem to be solved by the invention]

[0004] The problem to be solved by the present invention is to provide a highly accurate method for calculating skin age using an index that is highly correlated with actual age, a makeup method for improving skin age with cosmetics, and a method for evaluating the effect of cosmetics in improving skin age. [Means for solving the problem]

[0005] The inventors analyzed many indices that have traditionally been used to analyze skin conditions and found that some indices have a high correlation with the subject's actual age and some have a low correlation, and that the RGB B value of the forehead, the right cheek, and the left cheek in particular have a high correlation with the subject's actual age. They then used these indices to calculate the skin age with high accuracy, thereby completing the present invention.

[0006] The present invention encompasses the embodiments described below. Section 1. Obtaining a regression equation for calculating skin age based on a correlation between a variable including at least one of (i) the B value of the RGB of the forehead and the B value of the RGB of the cheeks, and (ii) the G value of the RGB of the cheeks of a face image and actual age; and A method for calculating a skin age, comprising: calculating a skin age of a subject based on the regression equation from variables including at least one of the subject's (i) RGB B value of the forehead and the RGB B value of the cheeks, and (ii) RGB G value of the cheeks. Section 2. 2. The method for calculating skin age according to item 1, wherein the variables include a B value of RGB of the forehead and a B value of RGB of the cheek. Section 3. Item 2. A method for calculating skin age according to item 1, wherein the variables include 80% or more of the features of the number of pores, the number of pore pixels, the number of wrinkles, the number of fine wrinkles, the R value of skin color, the B value of skin color, the G value of skin color, the average RGB of skin color, the area of ​​gloss, the color of gloss, redness, pigment changes other than redness, melanin, sebum, and porphyrin in each of the forehead, the back of the nose, the corners of the eyes, under the eyes, the cheeks, and the nasolabial folds, with the feature importance of the features including those of porphyrin having a positive value. Section 4. The variables include at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, as well as pixels of pores on the forehead, the number of pores on the forehead, wrinkles on the forehead, fine wrinkles on the forehead, RGB G value of the forehead, redness on the forehead, pigmentary changes other than redness on the forehead, melanin on the forehead, sebum on the forehead, pixels of pores on the dorsal nasal region, the number of pores on the dorsal nasal region, redness on the dorsal nasal region, pigmentary changes other than redness on the dorsal nasal region, wrinkles on the dorsal nasal region, RGB G value of the dorsal nasal region, gloss area on the dorsal nasal region, gloss color on the dorsal nasal region, melanin on the dorsal nasal region, sebum on the dorsal nasal region, pixels of pores on the dorsal nasal region, wrinkles on the dorsal nasal region, fine wrinkles on the dorsal nasal region, redness on the dorsal nasal region, RGB R value of the dorsal nasal region, RGB G value of ... Item 2. A method for calculating skin age according to Item 1, comprising 45 or more selected from the group consisting of porphyrin under the eyes, pixels of pores under the eyes, number of pores under the eyes, wrinkles under the eyes, fine wrinkles under the eyes, RGB R value under the eyes, RGB G value under the eyes, redness under the eyes, pigment changes other than redness under the eyes, melanin under the eyes, gloss area under the eyes, gloss color under the eyes, sebum under the eyes, porphyrin under the eyes, pixels of pores on the cheeks, number of pores on the cheeks, wrinkles on the cheeks, fine wrinkles on the cheeks, RGB R value of the cheeks, RGB G value of the cheeks, average skin color of the cheeks, gloss area on the cheeks, redness on the cheeks, pigment changes other than redness on the cheeks, melanin on the cheeks, sebum on the cheeks, porphyrin on the cheeks, wrinkles in the nasolabial folds, and fine wrinkles in the nasolabial folds. Section 5. Item 2. The method for calculating a skin age according to Item 1, further comprising: substituting into the regression equation all possible values ​​of at least one of the variables including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, and fixing the remaining variables to values ​​obtained for the subject; and calculating, from the multiple values ​​of the skin age obtained, a value that gives the lowest skin age. Section 6. A cosmetic method for improving skin aging, comprising at least one of the following steps (I) to (III): (I) applying a cosmetic product to the forehead, right cheek, and left cheek of a subject such that the skin age of the subject after application of the cosmetic product is lower than the skin age of the subject before application of the cosmetic product, the skin age being calculated by the method according to any one of items 1 to 4. (II) applying a cosmetic product to the forehead, right cheek, and left cheek of a subject such that the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks of the subject after applying the cosmetic product are all higher than the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks of the subject before applying the cosmetic product; (III) applying a cosmetic product to the right and left cheeks of a subject such that the G value of the RGB of the cheeks of the subject after applying the cosmetic product is lower than the G value of the RGB of the cheeks of the subject before applying the cosmetic product; The method includes: Section 7. Item 7. The method according to Item 6, comprising the step (I), wherein the cosmetic is applied so that a value obtained by subtracting the skin age of the subject before application of the cosmetic from the skin age of the subject after application of the cosmetic is -5 or less. Section 8. Item 7. The method according to item 6, comprising the step (II), wherein the cosmetic is applied such that the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks of the subject after application of the cosmetic are increased by 10 or more, respectively, compared with the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks of the subject before application of the cosmetic. Section 9. A method for evaluating the skin aging improvement effect of a cosmetic product, comprising: Calculating the skin age of the subject before and after use of the cosmetic product by the method described in any one of items 1 to 4; and A method comprising: evaluating a cosmetic product by comparing the skin age of the subject before use of the cosmetic product with the skin age of the subject after use of the cosmetic product. Section 10. A skin age estimation system for estimating a skin age from a facial image of a subject, comprising: a storage means for storing a regression equation, which is a calculation equation for skin age, generated based on a correlation between a variable including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, of a face image, and an actual age; a calculation means for calculating a skin age of a subject based on the regression equation from variables including at least one of the subject's (i) forehead RGB B value and cheek RGB B value, and (ii) cheek RGB G value; A skin age estimation system equipped with Section 11. Item 11. The skin age estimation system according to item 10, further comprising a skin age display means for displaying the skin age calculated by the calculation means on a display device. Effect of the Invention

[0007] According to the method for calculating skin age of the present invention, it is possible to calculate a skin age reflecting an actual age with high accuracy. Also, according to the makeup method for improving skin age of the present invention, it is possible to improve skin age simply and effectively. According to the method for evaluating the effect of cosmetics on improving skin age of the present invention, it is possible to evaluate the effect of cosmetics on improving skin age simply and more reliably. [Brief description of the drawings]

[0008] [Figure 1] 1 is a schematic diagram of a skin age estimation system according to an embodiment of the present invention. [Diagram 2] An example of the analysis results for the "pores" category displayed on the screen of the NeoVoir display device. (Left) Image of the subject's face overlaid with the distribution of pores, (top center) an illustration showing the number of pores, (center right) a bar graph comparing the number of pores in each part of the subject's body with the average number of pores in people of the same age, (bottom right) numerical data on the location, number and size of pores. [Diagram 3] Distribution of subjects who underwent NeoVoir skin analysis. Horizontal axis: age, vertical axis: number of people [Figure 4] Ten facial regions A to J determined by deep learning. A: forehead, B: nose bridge, C: right eye corner, D: left eye corner, E: right under eye, F: left under eye, G: right cheek, H: left cheek, I: right nasolabial fold, J: left nasolabial fold [Diagram 5] Importance of each item (feature) in the prediction model before improvement. The top 10 items are also shown. [Figure 6] The importance of each item (feature) according to the prediction model. The top 10 items are also shown. The top 10 items in the prediction model before improvement are also shown in the box on the right for comparison. [Figure 7]Graph of actual age and skin age calculated from the prediction model. (A) Predicted skin age vs. actual age, (B) Number of people vs. difference between predicted skin age and actual age ((Predicted skin age) - (Actual age)). [Figure 8] A graph showing the difference between predicted skin age and actual age (predicted skin age - actual age). Vertical axis: Number of people [Figure 9] Heat map showing the correlation between RGB value / skin color and age in each part of the body. (A) R value (B) G value (C) B value (D) Skin color (average value of RGB) (E) Reference: Examples of each color and RGB [Figure 10] SHAP value for each item. Items at the top have higher feature values. [Figure 11] A graph showing the difference between predicted skin age and actual age (predicted skin age - actual age). Vertical axis: Number of people [Figure 12] A graph showing the relationship between the subjects' cheek G value and their actual age. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] 1. How to calculate skin age The present invention provides a method for calculating skin age, which includes obtaining a regression equation for calculating skin age based on a correlation between a variable including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks of a face image, and actual age, and calculating the skin age of a subject based on the regression equation from a variable including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks of the subject.

[0010] The variables including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks of the face image, and the data on actual age are obtained from multiple subjects.

[0011] Hereinafter, the B value of RGB may be simply referred to as the "B value."

[0012] In this specification, "forehead" and "forehead" may be used interchangeably.

[0013] In this specification, the numerical value of a body part having left and right parts includes the numerical value of one of the parts, the separate numerical values ​​of both parts, and one variable calculated based on the separate numerical values ​​of both parts, and any numerical value may be used. For example, the B value of the cheek may use two B values ​​of the left and right cheeks, or the average value of the B values ​​of the left and right cheeks, but in terms of measurement accuracy and calculation efficiency, the average value of the B values ​​of the left and right cheeks is preferable. Similarly, the G value of the cheek may use two B values ​​of the left and right cheeks, or the average value of the B values ​​of the left and right cheeks, but in terms of measurement accuracy and calculation efficiency, the average value of the G values ​​of the left and right cheeks is preferable.

[0014] Surprisingly, the inventors found that the RGB B values ​​of the forehead, right cheek, and left cheek of the facial images of multiple subjects, i.e., the degree of blue in three specific areas, are inversely correlated with the actual ages of the multiple subjects, and are strongly related to each other. Also, the inventors found that the G values ​​of the cheeks of the facial images of multiple subjects, i.e., the degree of green in the cheeks, are positively correlated with the actual ages of the multiple subjects. Based on a regression equation created using these variables, the skin age of the subject can be calculated with high accuracy from the facial image of the subject.

[0015] The RGB B values ​​of the forehead and cheeks, and the RGB G values ​​of the cheeks of each of the face images of the multiple subjects can be obtained by any known method. For example, in the case of the RGB B values ​​of the forehead, the forehead area is set in the photographed face image of the subject by machine learning such as deep learning. Meanwhile, the image is separated into three components of RGB, and a specific area of ​​the image can be expressed by RGB values ​​(R, G, B). Therefore, the RGB B values ​​of each pixel in the selected forehead area can be calculated, and the average value within the area can be calculated. The B values ​​of the right cheek and the left cheek can also be calculated in the same way as the forehead.

[0016] The above regression equation is Equation (1):

[0017]

number

[0018] The coefficient of determination, R, is 2 Any regression equation may be used as long as the coefficient of determination R is 0.5 or more. 2 is preferably 0.6 or more, more preferably 0.65 or more, even more preferably 0.68 or more, and even more preferably 0.7 or more.

[0019] For example, multiple regression analysis with two or more explanatory variables is expressed by the following formula.

[0020] Y = A1X1+A2X2+A3X3+…A n X n +B In the formula, X1, X2, X3,...X n are explanatory variables, A1, A2, A3, …A n , B is a coefficient, and Y is the dependent variable.

[0021] Alternatively, the regression equation may be obtained using Lasso regression, Ridge regression, Elastic Net, Support Vector Regression, a decision tree model (e.g., XGBoost, LightGBM, CatBoost, etc.), or a machine learning tool provided on the Internet (e.g., Auto Gluon, URL: https: / / auto.gluon.ai / stable / index.html), etc.

[0022] In a regression equation in which a variable including at least one of (i) the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks, and (ii) the RGB G values ​​of the cheeks of a plurality of subjects is used as an explanatory variable, and the actual ages of a plurality of subjects are used as the objective variable, one or more other values ​​may be used as explanatory variables.

[0023] In one embodiment, the variables (explanatory variables) include the B value of the RGB of the forehead and the B value of the RGB of the cheeks.

[0024] In another embodiment, the variables (explanatory variables) include G values ​​of RGB of the cheeks.

[0025] In another embodiment, the variables (explanatory variables) include a B value of the RGB forehead, a B value of the RGB for cheeks, and a G value of the RGB for cheeks.

[0026] Preferably, the B value and G value of the RGB of the cheek are the average value of the B value and the average value of the G value of both cheeks, respectively.

[0027] In one embodiment, the above variables (explanatory variables) include 80% or more of the features with positive feature importance (Feature Importance), preferably the top 90% or more, and more preferably all of the features with positive feature importance (Feature Importance) among the features of the forehead, nose, eye corners, under the eyes, cheeks, and nasolabial folds, such as the number of pores, the number of pore pixels, the number of wrinkles, the number of fine wrinkles, the R value of skin color, the B value of skin color, the G value of skin color, the average RGB of skin color, the area of ​​gloss, the color of gloss, redness, pigment changes other than redness, melanin, sebum, and porphyrin. These variables also include at least one of (i) the B value of the RGB of the forehead and the B value of the RGB of the cheeks, and (ii) the G value of the RGB of the cheeks. By using a large number of explanatory variables with high feature importance in this way, it is possible to calculate the skin age with high accuracy.

[0028] Number of pores, number of pore pixels, number of wrinkles, number of fine wrinkles, R value of skin color, B value of skin color, G value of skin color, average RGB value of skin color, gloss area, gloss color, redness, pigment changes other than redness (In NeoVoir, areas of the skin that are not the usual color are considered to be pigment changes, and these are non-red areas. These are thought to represent blemishes that can be seen with polarized light.) Melanin, sebum, and porphyrin are all well-known indicators that are often used when analyzing skin condition, and although they are measured using NeoVoir (registered trademark) in the examples, they can be measured by any well-known method.

[0029] In one embodiment, the variables (explanatory variables) include at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, as well as pixels of pores on the forehead, the number of pores on the forehead, wrinkles on the forehead, fine wrinkles on the forehead, RGB G value of the forehead, redness on the forehead, pigmentary changes other than redness on the forehead, melanin on the forehead, sebum on the forehead, pixels of pores on the dorsal nasal, the number of pores on the dorsal nasal, redness on the dorsal nasal, pigmentary changes other than redness on the dorsal nasal, wrinkles on the dorsal nasal, RGB G value of the dorsal nasal, gloss area on the dorsal nasal, gloss color on the dorsal nasal, melanin on the dorsal nasal, sebum on the dorsal nasal, pixels of pores on the dorsal nasal, the number of pores on the dorsal nasal, wrinkles on the dorsal nasal, fine wrinkles on the dorsal nasal, redness on the dorsal nasal, RGB R value of the dorsal nasal, RGB G value of the dorsal nasal, gloss color on the dorsal nasal, The present invention includes 45 or more selected from the group consisting of melanin at the corner of the eye, sebum at the corner of the eye, porphyrin at the corner of the eye, pixels of pores under the eyes, number of pores under the eyes, wrinkles under the eyes, fine wrinkles under the eyes, RGB R value under the eyes, RGB G value under the eyes, redness under the eyes, pigment changes other than redness under the eyes, melanin under the eyes, gloss area under the eyes, gloss color under the eyes, sebum under the eyes, porphyrin under the eyes, pixels of pores on the cheeks, number of pores on the cheeks, wrinkles on the cheeks, fine wrinkles on the cheeks, RGB R value of the cheeks, RGB G value of the cheeks, average skin color of the cheeks, gloss area on the cheeks, redness on the cheeks, pigment changes other than redness on the cheeks, melanin on the cheeks, sebum on the cheeks, porphyrin on the cheeks, wrinkles in the nasolabial folds, and fine wrinkles in the nasolabial folds. The latter option is more preferably 50 or more, and more preferably all of them are included. By using a large number of explanatory variables with high feature importance in this way, it is possible to calculate the skin age with high accuracy.

[0030] In one embodiment, the variables (explanatory variables) include at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, and the pixels of L-sized pores on the forehead, L-sized melanin on the corners of the eyes, L-sized melanin on the dorsum of the nose, wrinkles in the nasolabial fold, sebum on the dorsum of the nose, the number of L-sized pores on the forehead, wrinkles on the forehead, glossy area on the cheeks, S-sized melanin on the forehead, pigmentary changes other than redness under the eyes, S-sized melanin on the dorsum of the nose, glossy area on the dorsum of the nose, fine wrinkles in the nasolabial fold, L-sized melanin on the cheeks, the number of L-sized pores on the cheeks, wrinkles under the eyes, sebum on the forehead, L-sized pore pixels, large melanin on forehead, fine wrinkles on cheeks, redness on cheeks, small pore pixels under eyes, number of small pores on corner of eyes, large pore pixels under eyes, fine wrinkles on corner of eyes, porphyrin on cheeks, large melanin under eyes, RGB G value on cheeks, gloss color under eyes, medium pore pixels on corner of eyes, RGB R value on corner of eyes, wrinkles on corner of eyes, medium melanin on cheeks, redness on back of nose, large pore pixels on corner of eyes, number of small pores under eyes, wrinkles on back of nose, wrinkles on cheeks, medium pore pixels on cheeks, RGB G value on back of nose, skin on cheeks Oil, cheek skin color, RGB G value under the eyes, medium-sized melanin at the corner of the eye, redness under the eyes, pixels of medium-sized pores on the forehead, number of medium-sized pores at the corner of the eye, medium-sized melanin on the back of the nose, pixels of small-sized pores on the cheeks, pigment changes other than redness on the forehead, pixels of medium-sized pores under the eyes, RGB R value on the cheeks, medium-sized melanin under the eyes, small-sized melanin under the eyes, small-sized melanin at the corner of the eye, gloss area under the eyes, number of pores on the forehead, number of large-sized pores under the eyes, medium-sized melanin on the forehead, gloss color at the corner of the eye, pixels of small-sized pores at the corner of the eye, medium-sized hair on the cheeks and 70 or more selected from the group consisting of: number of pores on the forehead, fine wrinkles under the eyes, redness on the forehead, redness at the corner of the eyes, number of small pores on the bridge of the nose, porphyrin at the corner of the eyes, number of small pores on the forehead, fine wrinkles on the forehead, sebum under the eyes, pigment changes other than redness on the bridge of the nose, pixels of small pores on the forehead, pixels of medium pores on the bridge of the nose, porphyrin under the eyes, RGB R value under the eyes, RGB G value of the corner of the eyes, number of medium pores on the bridge of the nose, number of medium pores under the eyes, number of large pores on the bridge of the nose, sebum at the corner of the eyes, pigment changes other than redness on the cheeks, gloss color of the bridge of the nose, and RGB G value of the forehead.The latter option is more preferably 75 or more, more preferably 80 or more, and more preferably includes all of them. By using a large number of explanatory variables having high feature importance in this way, it is possible to calculate the skin age with high accuracy.

[0031] In any of the above embodiments, the coefficient of determination R 2 A person skilled in the art can select an appropriate number and type of explanatory variables within the scope of ordinary skill so as to obtain a regression equation with a σ of 0.5 or more.

[0032] The method of calculating skin age of the present invention may further include substituting into the regression equation all possible values ​​of at least one of variables including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, fixing the remaining variables to the values ​​obtained for the subject, and calculating the lowest skin age value from the multiple skin age values ​​obtained.

[0033] With this configuration, at least one of the RGB B value of the forehead and the RGB B value of the cheeks and the RGB G value of the cheeks that give the subject's skin age the lowest can be obtained, and the value can be used as an index for improving skin age. For example, when applying cosmetics, the type and / or amount of the cosmetics can be adjusted so that the skin age approaches the lowest value.

[0034] Furthermore, using a similar method, it is possible to vary other variables within their possible range of values ​​while fixing the other variables, inputting both into a regression equation, and comparing the resulting skin age values ​​to determine the value that gives the lowest skin age for that variable.

[0035] In one embodiment, at least one of the RGB B value of the forehead and the RGB B value of the cheek and the RGB G value of the cheek that give the lowest skin age of the subject includes the RGB B value of the forehead and the RGB B value of the cheek.

[0036] In another embodiment, at least one of the RGB B value of the forehead and the RGB B value of the cheeks, which give the lowest skin age of the subject, includes the RGB G value of the cheeks.

[0037] In another embodiment, the skin age of the subject is lowest when the RGB B value of the forehead, the RGB B value of the cheek, or the RGB G value of the cheek is at least one of the RGB B value of the forehead, the RGB B value of the cheek, and the RGB G value of the cheek.

[0038] Preferably, the B value and G value of the RGB of the cheek are the average value of the B value and the average value of the G value of both cheeks, respectively.

[0039] The gender of the subjects may be either male or female, but it is preferable to match the gender of the subjects (for example, only female) in order to improve the accuracy of the calculation of the skin age. It is preferable that the gender of the subject whose skin age is to be calculated is the same as the gender of the subjects in order to improve the accuracy of the calculation of the skin age.

[0040] The subjects may be of the same age group (e.g., 20s, 30s, 40s, 50s, 60s, or 70s) or may be of different age groups (e.g., 20s, 30s, 40s, 50s, 60s, and 70s). The number of subjects in each age group is preferably 10 or more, and more preferably 20 or more, in order to improve the accuracy of skin age calculation.

[0041] The actual age of the subject for calculating the skin age is not particularly limited, and may be, for example, in the teens to nineties, twenties to eighties, or thirties to seventies.

[0042] 2. Makeup techniques to improve skin aging The present invention relates to a cosmetic method for improving skin aging, the method comprising at least one of the following steps (I) to (III): (I) applying a cosmetic product to the forehead, right cheek, and left cheek of a subject such that the skin age of the subject after applying the cosmetic product is lower than the skin age of the subject before applying the cosmetic product, the skin age being calculated by the method for calculating the skin age of the subject according to the above-mentioned 1. (II) applying the cosmetic product to the forehead, right cheek, and left cheek of the subject such that the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks of the subject after applying the cosmetic product are all higher than the RGB B values ​​of the cheeks of the subject before applying the cosmetic product; (III) applying a cosmetic product to the right and left cheeks of a subject such that the G value of the RGB of the cheeks of the subject after applying the cosmetic product is lower than the G value of the RGB of the cheeks of the subject before applying the cosmetic product; The present invention provides a method comprising:

[0043] In one embodiment, the cosmetic method for improving skin age is a method comprising step (I). Skin age can be reduced, for example, by applying a cosmetic comprising a blue pigment to the forehead and cheeks of a subject as described above. In another embodiment, the cosmetic method for improving skin age is a method comprising step (II). In another embodiment, the cosmetic method for improving skin age is a method comprising step (III). In another embodiment, the cosmetic method for improving skin age is a method comprising steps (II) and (III). In another embodiment, the cosmetic method for improving skin age is a method comprising steps (I) to (III).

[0044] Since the B value of RGB indicates blue, the B value of each of the forehead, right cheek, and left cheek can be easily adjusted by makeup. In other words, by applying cosmetics to at least the forehead, right cheek, and left cheek to increase the B value of each area, the skin age can be easily reduced compared to before the application of the cosmetics. Such improvement in skin age by adjusting the B value is also more rapid than improvement in skin age by improving the opening of pores or wrinkles. Examples of such cosmetics include a makeup base containing a blue pigment, a concealer containing a blue pigment, and a foundation containing a blue pigment. The blue pigment may be any blue pigment that can be used in cosmetics, and includes, but is not limited to, Gunjo, Gunjo Violet, Konjo, Blue No. 1 (Brilliant Blue FCF), Blue No. 2 (Indigo Carmine), Gardenia Blue Liquid, and the like. Applying cosmetics includes, for example, adjusting the type and / or amount of cosmetics.

[0045] When applying the cosmetic product to the subject's face, the cosmetic product may be applied only to the right and left cheeks, or only to the forehead, right and left cheeks, or to these and other areas. The cosmetic product may also be applied to substantially the entire face, including the forehead, right and left cheeks.

[0046] Since the G value of RGB indicates green, the G value of the forehead can be easily adjusted by makeup. In other words, by applying cosmetics to at least the forehead to lower the G value of the forehead, it is possible to easily lower the skin age compared to before the application of the cosmetics. For example, it is possible to apply a color corresponding to B, or to apply red, which is the complementary color of G. Cosmetics containing blue pigments or cosmetics containing red pigments are effective in lowering the G value of the forehead, but the above-mentioned cosmetics containing blue pigments are particularly effective in increasing the B value of the cheeks and lowering the G value.

[0047] Cosmetics containing a red pigment include a makeup base containing a red pigment, a concealer containing a red pigment, a foundation containing a red pigment, etc. The red pigment may be any red pigment that can be used in cosmetics, and includes, but is not limited to, Red No. 104, Red No. 106, Red No. 201, Red No. 202, Red No. 205, Red No. 226, Red No. 228, Red No. 3, carmine, red iron oxide (iron (III)), etc.

[0048] According to the above-mentioned makeup method, skin aging can be improved simply and effectively with cosmetics.

[0049] In one embodiment, the method includes step (I), in which the cosmetic is applied so that the skin age of the subject after application of the cosmetic is subtracted from the skin age of the subject before application of the cosmetic is equal to or less than -5. Applying the cosmetic includes, for example, adjusting the type and / or amount of the cosmetic. According to this configuration, the cosmetic is applied so that the effect of improving skin age can be confirmed, and therefore the skin age can be reliably improved.

[0050] In one embodiment, the method includes step (II), in which the cosmetic is applied such that the RGB B value of each of the regions of the right cheek, left cheek, and forehead of the subject after application of the cosmetic increases by 10 or more, respectively, compared with the B value of the right cheek, left cheek, and forehead of the subject before application of the cosmetic. Applying the cosmetic includes, for example, adjusting the type and / or amount of the cosmetic. With this configuration, the skin age is reduced by about 5 years, and therefore the skin age can be reliably improved.

[0051] 3. Method for evaluating the effect of cosmetics on improving skin aging The present invention relates to a method for evaluating the effect of cosmetics on improving skin age, comprising: calculating the skin age of a subject before and after use of the cosmetics by the method for calculating skin age described in 1 above; The method includes evaluating a cosmetic product by comparing the skin age of the subject before use of the cosmetic product with the skin age of the subject after use of the cosmetic product.

[0052] The above-described evaluation method of the present invention is capable of capturing minute changes in skin aging caused by application of cosmetics, and is therefore useful for evaluating the effect of cosmetics in improving skin aging.

[0053] In one embodiment, the subject is a single subject, and the skin age of the subject before use of a certain cosmetic product is compared with the skin age of the subject after use of the cosmetic product. If the skin age of the subject after use of the cosmetic product is lower than the skin age of the subject before use of the cosmetic product, this indicates that the cosmetic product has a skin age-improving effect on the subject.

[0054] In another embodiment, the subject is a plurality of subjects, and the average skin age of the plurality of subjects before use of a certain cosmetic product is compared with the average skin age of the plurality of subjects after use of the cosmetic product; if the average skin age of the plurality of subjects after use of the cosmetic product is lower than the average skin age of the plurality of subjects before use of the cosmetic product, this indicates that the cosmetic product has an effect of improving skin age.

[0055] 4. Skin Age Estimation System Next, a skin age estimation system for carrying out the skin age calculation method of the present invention will be described. As shown in Fig. 1, the skin age estimation system 100 for estimating the skin age from the face image of the subject according to the embodiment of the present invention includes a storage means 112 for storing a regression equation, which is a calculation equation for the skin age, generated based on a correlation between a variable including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks and (ii) the RGB G value of the cheeks of the face image and the actual age, and a skin age calculation means 122 for calculating the skin age of the subject based on the regression equation from the RGB B value of the forehead and the RGB B value of the cheeks of the face image of the subject.

[0056] The variables including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks of the face image are as explained in "1. Method of calculating skin age."

[0057] The skin age estimation system 100 includes an imaging device 102 for capturing a facial image of a subject, an RVB value generating means 104 for generating an RVB value based on facial image data captured by the imaging device 102, a memory device 100 such as a ROM having a memory means 112, a processing device 120 (e.g., a control unit such as a Central Processing Unit (CPU)) having a regression model generating means 122, a skin age calculation means 124, a display screen generating means 126, and a skin age display means 128, and a display device 130 (e.g., a display).

[0058] Facial images of a plurality of subjects captured by the imaging device 102 are decomposed into RVB by the RVB value generating means 104, which generates B values ​​and / or G values ​​of each region such as the forehead, right cheek, and left cheek. The generated B values ​​and / or G values ​​of the forehead, right cheek, and left cheek are sent to the storage device 110 and stored therein. Data on the actual ages of the plurality of subjects is also stored in the storage device 110. The generation of the regression equation is executed by, for example, a program stored in the recording medium 114 of the storage device 110. The regression model generating means 122 generates a regression equation from a plurality of variables including at least one of (i) the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks, and (ii) the RGB G values ​​of the cheeks, and the actual ages of the plurality of subjects.

[0059] Next, the skin age calculation means 124 calculates the skin age of the subject, whose skin age is to be calculated, based on the above regression equation from variables including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, of the facial image of the subject received from the RVB value generation means 104.

[0060] Next, the display screen generating means 126 generates a screen displaying the skin age from the skin age calculated by the skin age calculating means 124, and the skin age display means 128 causes the display device 130 to display the screen displaying the skin age created by the display screen generating means 126.

[0061] With this configuration, the skin age of a subject can be calculated simply by capturing a facial image of the subject whose skin age is to be calculated.

[0062] Optionally, the processing device 120 may further include a minimum skin age calculation means 125. The minimum skin age calculation means 125 substitutes all possible values ​​of at least one of (i) the B value of the RGB of the forehead and the B value of the RGB of the cheeks, and (ii) the G value of the RGB of the cheeks, stored in the storage device 110, of the subject whose skin age is to be calculated, into the regression equation, fixes the remaining variables to the values ​​obtained for the subject, and calculates the value that gives the lowest skin age from the multiple skin age values ​​obtained for the subject.

[0063] A display screen generating means 126 generates a screen displaying the skin age from the minimum skin age calculated by the minimum skin age calculating means 125, and a skin age display means 128 causes the display device 130 to display the screen displaying the minimum skin age created by the display screen generating means 126.

[0064] The disclosures of all patent applications and publications cited herein are hereby incorporated by reference in their entireties.

[0065] The following examples are intended for illustrative purposes only and are not intended to limit the scope of the present invention in any way. EXAMPLES

[0066] Example 1 Analysis of skin condition based on face images 1. Explanation of analysis method using NeoVoir (registered trademark) (Ci:Labo Co., Ltd.) NeoVoir is a skin analyzer that uses a long-life LED light source and performs a composite analysis using four types of light (general light, glossy light, polarized light, and ultraviolet light). It is a machine that can take images from three sides (front, left, right, and side) and evaluate skin changes in areas recognized by deep learning using quantitative data. Details are described in Example 2 (2) Materials.

[0067] Figure 2 shows a computer display screen that displays a face image taken using NeoVoir and the quantitative data obtained. Although NeoVoir allows comparisons with the average values ​​of each item by age group, skin age is calculated by adding and subtracting the influence of each item to the subject's actual age (details are unknown), and skin age has not been estimated based only on the measured quantitative data without using the subject's actual age. Therefore, we conducted a more detailed analysis of the relationship between the quantitative data measured by NeoVoir and skin age.

[0068] 2. Examination of skin analysis data using NeoVoir (1) Target The subjects were 2,543 people (2,439 women / 104 men, ages 17 to 71) who underwent NeoVoir skin analysis (Figure 3).

[0069] (2) Material NeoVoir (registered trademark) is a device consisting of a skin image capture device and an analysis application. This device captures facial images using four types of light: general light, ultraviolet light, and two types of polarized light. Deep learning technology is applied to the captured images to automatically draw the following areas (Figure 4): forehead (A), nose back (B), corners of the eyes (right: C, left: D), under the eyes (right: E, left: F), cheeks (right: G, left: H), and nasolabial folds (right: I, left: J). Using polarized light, areas that are perceived as redness, which is different from normal skin color, are defined as redness, and areas that are not redness are defined as notred. In this study, the Shadow Method was used to evaluate pores and wrinkles. The Shadow Method is a method in which both sides and above and below the eyebrows are indirectly illuminated with a general light source to measure the shadow areas. In this method, green (G) and brown (B), which are associated with black, are excluded from RGB, and only the red (R) value is measured. Since black is not measured, eyebrows and eyelashes are excluded, and shadows caused by pores can be identified by color. The number of shadows was measured and defined as pores. In addition, using the same shadow method, shadows thought to be caused by wrinkles were also measured and defined as wrinkles. By taking pictures with horizontally polarized light that does not suppress light reflection, it is possible to detect shadows that are not visible with normal light sources. Of these, shadows thought to be caused by fine wrinkles were measured and defined as fine wrinkles (fwrinkles). Skin color was expressed by the RGB values ​​of each area taken with standard light, and the red, green, and blue values ​​in the RGB color model were defined as r, g, and b, respectively. The color and area of ​​the analysis area were extracted from only the high-brightness areas (gloss) from polarized light that is easy to detect on a flat surface, and were defined as lcolor and larea, respectively. Melanin was defined as the dark areas measured using only the shadow color tone from an ultraviolet light source in which melanin is easily detected. For pores and melanin, the measurements were divided by size, with 20-70 pixels being small (s), 71-120 pixels being medium (m), and 121 pixels or more being large (l). The sizes l, m, and s correspond to L, M, and S sizes. Sebum was identified by extracting only white and blue fluorescence using the high brightness gradation of a UV light source. Of the extracted sebum, only orange and red were extracted and identified as porphyrin.For example, porelpx refers to the number of pixels of L-sized pores, melaninl refers to the number of L-sized melanins, and porelcnt refers to the number of L-sized pores.

[0070] The following numerical data were measured for each region of the AJ: Red notred Pigment change other than redness larea Glossy area lcolor Gloss color skincolor Average RGB value of skin color b Blue value (B value) g Green value (G value) r Red value (R value) melanin (s, m, l) pore cnt Number of pores (s, m, l) pore px Number of pore pixels (s, m, l) porphyrin sebum fwrinkle Fine wrinkles wrinkle wrinkles

[0071] (3) Data analysis Numerical data were exported from NeoVoir® as CSV files, which were analyzed using the Python 3.8 programming language. All data were randomly assigned to the training set and validation set in a ratio of 8:2 during model construction to avoid overfitting and poor generalization performance. Furthermore, to prevent poor model performance due to excess variables, the average of the left and right sides was calculated for symmetrical regions, and calculations were performed based on this average. Since the algorithm used can only process numerical data, gender was converted to a numerical value (0 for male and 1 for female) before analysis. The coefficient of determination (R 2 ) was used to measure the goodness of fit of a model by quantifying the reduction in error compared to the mean ground truth of the prediction. Values ​​closer to 1 indicate a better model fit.

[0072] (4) Results The trends in the relationship between gender and age varied depending on the item, as shown below. The number of S pores on the corners of the eyes and cheeks increases with age. On the other hand, the number of S pores under the eyes and on the nose is almost inconsistent, and on the forehead it decreases with age. The number of M / L pores increases gradually with age except on the forehead, where it remains almost constant. Wrinkle / fwrinkle tended to increase with age in all areas, and there was little difference between the sexes. Skin color and RBG decreased with age. Men had lower values. Larea decreased with age in the cheeks. Overall, men had lower values, but there was little difference between the sexes in the nose. Overall, lcolor decreased with age. The incidence of brown pigmented lesions (notred) increased with age, with little difference between men and women, although on the forehead it was slightly more prevalent in men. - Red pigmented lesions showed no trend with age. Melanin generally increases with age, with little difference between the sexes. Melanin in the S under the eyes showed little change with age. Sebum production decreased with age on the forehead, nose, and cheeks. - Porphyrins showed no clear trend. To account for the possibility that the model may not have been able to learn as well about men due to the small number of male samples, we generated regression models using only the female data.

[0073] In the following Examples 2-7, for bilaterally symmetrical regions, the measured values ​​of the items in each of the left and right regions were used.

[0074] Example 2. Selection of regression model method We used different models, including multiple regression model, lasso regression, ridge regression, elastic net, support vector regression, XGBoost, LightGBM, and CatBoost, and all of them had a coefficient of determination of 0.5 or higher (0.626, 0.634, 0.637, 0.637, 0.617, 0.607, 0.607, and 0.622, respectively). To further improve accuracy, we generated a model using the machine learning automation library Auto Gluon. The coefficient of determination rose to 0.695 (Figure 5). Furthermore, we deleted items with negative feature importance and generated a model using Auto Gluon again, which resulted in a coefficient of determination of 0.703, and we adopted this model (Figure 6). result FIG. 7(A) is a graph showing the relationship between actual age (vertical axis) and skin age (horizontal axis) calculated from the prediction model. When this model was used, a high correlation was obtained between actual age and skin age calculated from the prediction model. In other words, when this model is used, it is possible to calculate a skin age that is highly correlated with actual age with high accuracy from data obtained from the subject's facial image, even without using data on the subject's actual age. FIG. 7(B) is a graph showing the relationship between the number of people (vertical axis) and the difference between the predicted skin age and actual age [difference between (predicted skin age) - (actual age)] (horizontal axis). The difference between the predicted skin age and actual age was distributed almost symmetrically around 0.

[0075] Example 3 Simulation of skin aging changes when topical agents are applied method A paper (Je Hyuk Lee, et al. Biomedical, January, 2021, Volume 33, 2, pp 25652-25658) evaluating the use of NeoVoir (Note: English name Mark-Vu (registered trademark)) before and after the use of Triple Combination Cream (TCC) (retinoid + HQ + steroid) and a microneedle patch reported that pigment changes improved by 20.46% and melanin by 16.6%. Therefore, a simulation data file was created by reducing notred by 20% and melanin by 16% in the female data file of Example 1, and in the created skin age prediction model, the skin age y_pred_all calculated from the original data and y_pred_simulation calculated from the simulation data were examined. result As shown in FIG. 8, the average difference between (predicted skin age) and (actual age) was -1.287 years, confirming that the application of topical agents can also change skin age.

[0076] Example 4: Search for an index suitable for estimating skin age method Using the data from Example 1, correlation coefficients were calculated between each color (RGB and their average, skincolor) in each body part (the letter at the end of the figure indicates the body part; A: forehead, B: nose, C: corner of the right eye, D: corner of the left eye, E: under the right eye, F: under the left eye, G: right cheek, H: left cheek) and between these colors and age. result Figures 9 (A) to (D) are heat maps of correlation coefficients. For all colors (R, B, G, and skincolor), cheeks (G, H) had the largest absolute correlation coefficient with age. In particular, B had the largest negative correlation coefficient with age among all items. 10 shows the top 20 items of SHAP values ​​(Shapley values) of explanatory variables in the model. The names of the measurement items are as explained in Example 1. The SHAP value is calculated by the following calculation method.

[0077]

number

[0078] It was found that the b values ​​of the cheeks and forehead (bG, bH, bA) have a significant impact on age prediction.

[0079] Example 5 Simulation of skin aging changes when a subject applies a makeup base method When a colored makeup base containing a blue pigment was applied to both cheeks and forehead of a subject and NeoVoir was photographed before and after, the b value changed by about 10. For this reason, in the data file of this study, a file of data in which bA, bG, and bH were increased by +10 was created as a simulation data file. In the skin age prediction model created in Example 2, the skin age y_pred_all calculated from the original data and y_pred_simulation calculated from the simulation data were examined. result As shown in Figure 11, the average difference between the predicted skin age and the actual age was -5.489 years, demonstrating that applying makeup to increase the B value of the forehead and cheeks can definitely make the skin look younger. In this way, it is possible to use "makeup" to improve skin age based on scientific theory.

[0080] Example 6: Calculation model for B value that gives the youngest skin age The numerical data of a subject was substituted unchanged for all items other than the B value of the right cheek into the skin age measurement model created in Example 2, and the B value of the right cheek was substituted successively with 256 possible values ​​(0-255). The B value of the right cheek which gave the lowest skin age in the skin age measurement model was determined. The execution result showed that a B value of 255 predicted the lowest (youngest) skin age, as shown in the program below. Similarly, it is possible to determine the B value of the left cheek, the B value of the forehead, or the combination of the B value of the right cheek / the B value of the left cheek / the B value of the forehead that results in the youngest skin age.

[0081]

number

[0082] Example 7 Correlation between cheek G value and age method Using the female data from the data file of Example 1, the relationship between the average RGB G value of both cheeks of the subjects and their actual age was examined. result As shown in Figure 12, the older the actual age, the stronger the green color of the cheeks. In the multiple regression analysis, the partial regression coefficient of gG was 0.8987, indicating that there is a positive correlation between the average G value of both cheeks and age. [Explanation of symbols]

[0083] 100: skin age estimation system, 112: storage means, 122: skin age calculation means, 128: skin age display means, 130: display device.

Claims

1. Obtaining a regression equation for calculating skin age based on a correlation between a variable including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheek, and (ii) the RGB G value of the cheek of the face image and actual age; and A method for calculating a skin age, comprising: calculating the skin age of a subject based on the regression equation from variables including at least one of the subject's (i) RGB B values ​​of the forehead and RGB B values ​​of the cheeks, and (ii) RGB G values ​​of the cheeks.

2. The method for calculating skin age according to claim 1 , wherein the variables include a B value of RGB of the forehead and a B value of RGB of the cheek.

3. 2. The method for calculating skin age according to claim 1, wherein the variables include 80% or more of the features of the number of pores, the number of pore pixels, the number of wrinkles, the number of fine wrinkles, the R value of skin color, the B value of skin color, the G value of skin color, the average RGB values ​​of skin color, the area of ​​gloss, the color of gloss, redness, pigment changes other than redness, melanin, sebum, and porphyrin in each of the forehead, the back of the nose, the corners of the eyes, under the eyes, the cheeks, and the nasolabial folds, with the feature importance of the features including positive values.

4. The variables include at least one of (i) RGB B values ​​of the forehead and RGB B values ​​of the cheeks, and (ii) RGB G values ​​of the cheeks, as well as pixels of pores on the forehead, number of pores on the forehead, wrinkles on the forehead, fine wrinkles on the forehead, RGB G values ​​of the forehead, redness on the forehead, pigmentary changes other than redness on the forehead, melanin on the forehead, sebum on the forehead, pixels of pores on the dorsal nasal, number of pores on the dorsal nasal, redness on the dorsal nasal, pigmentary changes other than redness on the dorsal nasal, wrinkles on the dorsal nasal, RGB G values ​​of the dorsal nasal, gloss area on the dorsal nasal, gloss color on the dorsal nasal, melanin on the dorsal nasal, sebum on the dorsal nasal, pixels of pores on the dorsal nasal, number of pores on the dorsal nasal, wrinkles on the dorsal nasal, fine wrinkles on the dorsal nasal, redness on the dorsal nasal, RGB R values ​​of the dorsal nasal, RGB G values ​​of the dorsal nasal, gloss color on the dorsal nasal, 2. The method of claim 1, further comprising 45 or more selected from the group consisting of porphyrin, pixels of pores under the eyes, number of pores under the eyes, wrinkles under the eyes, fine wrinkles under the eyes, RGB R value under the eyes, RGB G value under the eyes, redness under the eyes, pigment changes other than redness under the eyes, melanin under the eyes, gloss area under the eyes, gloss color under the eyes, sebum under the eyes, porphyrin under the eyes, pixels of pores on the cheeks, number of pores on the cheeks, wrinkles on the cheeks, fine wrinkles on the cheeks, RGB R value of the cheeks, RGB G value of the cheeks, average skin color of the cheeks, gloss area on the cheeks, redness on the cheeks, pigment changes other than redness on the cheeks, melanin on the cheeks, sebum on the cheeks, porphyrin on the cheeks, wrinkles in the nasolabial folds, and fine wrinkles in the nasolabial folds.

5. 2. The method for calculating skin age according to claim 1, further comprising: substituting into the regression equation each of all possible values ​​of at least one of variables including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, fixing the remaining variables to the numerical values ​​obtained for the subject, and calculating the value that gives the lowest skin age from the multiple skin age values ​​obtained.

6. A cosmetic method for improving skin aging, comprising at least one of the following steps (I) to (III): (I) applying a cosmetic product to the forehead, right cheek, and left cheek of a subject such that the skin age of the subject after application of the cosmetic product is lower than the skin age of the subject before application of the cosmetic product, the skin age being calculated by the method according to any one of claims 1 to 4. (II) applying the cosmetic product to the forehead, right cheek, and left cheek of the subject such that the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks of the subject after applying the cosmetic product are all higher than the RGB B values ​​of the forehead and the RGB B values ​​of the cheeks of the subject before applying the cosmetic product; (III) applying the cosmetic to the right and left cheeks of the subject such that the G value of the RGB of the cheeks of the subject after applying the cosmetic is lower than the G value of the RGB of the cheeks of the subject before applying the cosmetic; The method includes:

7. The method according to claim 6, comprising the step (I), wherein the cosmetic is applied so that a value obtained by subtracting the skin age of the subject before applying the cosmetic from the skin age of the subject after applying the cosmetic is -5 or less.

8. 7. The method according to claim 6, comprising step (II), wherein the cosmetic is applied such that the RGB B values ​​of the subject's forehead and cheeks after application of the cosmetic are increased by 10 or more, respectively, compared to the RGB B values ​​of the subject's forehead and cheeks before application of the cosmetic.

9. A method for evaluating the skin aging improvement effect of a cosmetic product, comprising: Calculating the skin age of a subject before and after use of a cosmetic product by the method according to any one of claims 1 to 4; and A method comprising: evaluating a cosmetic product by comparing the skin age of the subject before use of the cosmetic product with the skin age of the subject after use of the cosmetic product.

10. A skin age estimation system for estimating a skin age from a facial image of a subject, comprising: a storage means for storing a regression equation, which is a calculation equation for skin age, generated based on a correlation between a variable including at least one of (i) the RGB B value of the forehead and the RGB B value of the cheeks, and (ii) the RGB G value of the cheeks, of a face image, and an actual age; a calculation means for calculating a skin age of a subject based on the regression equation from variables including at least one of the RGB B value of the forehead and the RGB B value of the cheek of the subject (i) and the RGB G value of the cheek of the subject (ii); A skin age estimation system equipped with

11. The skin age estimation system according to claim 10 , further comprising a skin age display means for displaying the skin age calculated by the calculation means on a display device.

Images