JP7520679B2 - Method for estimating lively face degree, device for estimating lively face degree, and program for estimating lively face degree - Google Patents

Description

Article 30, Paragraph 2 of the Patent Act applied. On July 30, 2020, at the request of Pola Chemical Industries Co., Ltd., Pola Orbis Holdings Co., Ltd. published it on its website. <URL: http://www. pola-rm. co. jp/pdf/release_20200730_02.pdf>

The present invention relates to a method for estimating the degree of liveliness of a face, a device for estimating the degree of liveliness of a face, and a program for estimating the degree of liveliness of a face.

Having a "lively face" is one of the important evaluations for everyone. In particular, evaluation of the degree of "lively face" is extremely important in various fields such as esthetics and cosmetics.

Conventionally, there are known surveillance and recognition systems that input face images into a computer, recognize age groups, gender, and facial characteristics, and estimate a person. Patent documents 1 to 3 disclose age estimation systems that use image processing values such as edge processing and binarization (e.g., Patent documents 1 and 2), and person attribute estimation technology that uses feature quantities (position, shade, number of wrinkles) of feature points such as the eyes, nose, and mouth (e.g., Patent document 3).

There is also known technology that focuses on the fact that the texture characteristics of an image change depending on spatial periodicity (spatial frequency). As a technology that focuses on spatial frequency, Patent Document 4 discloses a method for creating simulation images of skin with various textures by changing the spatial frequency.

The applicant has also disclosed a technique for analyzing spatial frequency patterns when makeup is applied and evaluating the effect of makeup on the "three-dimensional appearance of the face" (for example, Patent Document 5).

JP 2009-086901 A JP 2009-271885 A JP 2008-282089 A JP 2004-283357 A JP 2004-272849 A

In the light of the above-mentioned prior art, the present inventors have conducted intensive research and have found that there is a high correlation between the intensity of spatial frequencies obtained from face image data and the degree of facial liveliness.
In other words, an object of the present invention is to provide a novel technique for estimating the degree of "lively face" with high accuracy and good reproducibility.

The present invention, which solves the above problem, is an estimation method that includes estimating the degree of livelyness of a subject's face from facial image data of the subject, based on the correlation between the intensity of spatial frequencies obtained from the facial image data and the degree of livelyness of the face.
As will be shown in the examples below, the intensity of spatial frequencies has a high correlation with a lively face.
Therefore, according to the estimation method of the present invention, it is possible to estimate the degree of facial liveliness of a subject from the facial image data of the subject with high accuracy and good reproducibility.

A preferred embodiment of the present invention includes a spatial frequency calculation step of calculating the intensity of a spatial frequency obtained by using data obtained from any channel of RGB color channel data of the color face image data, either alone or in combination;
an evaluation step of evaluating the lively face of the subject by applying the calculated spatial frequency intensity to a formula or model that indicates a correlation between the spatial frequency intensity obtained from face image data and the lively face level, the formula or model being prepared in advance;
The estimation method includes:

In a preferred embodiment of the present invention, the formula is a multiple regression formula obtained by multiple regression analysis or a prediction formula obtained by PLS.
By using a multiple regression equation or a prediction equation, it is possible to estimate the degree of facial liveliness of a subject from the facial image data of the subject with high accuracy and good reproducibility.

In a preferred embodiment of the present invention, the degree of facial vitality includes at least one of the following: brightness of the cheeks, dullness under the eyes, dullness of the cheeks, plumpness of the cheeks, complexion, firmness, yellowing of the entire face, roughness of the skin, and gloss of the skin.

In a preferred embodiment of the present invention, the facial image data is facial image data of a facial region or the entire face.

In a preferred embodiment of the present invention, the facial image data is color facial image data, and the spatial frequency is a spatial frequency obtained from at least one of the color channel data of RGB of the color facial image data.

In a preferred embodiment of the present invention, the spatial frequency is a spatial frequency obtained from a plurality of color channel data out of RGB color channel data of a face image.
By combining and using the spatial frequency intensities obtained from multiple color channel data out of the RGB color channel data of a facial image, it is possible to estimate (evaluate) the degree of liveliness of a subject's face from the facial image data of the subject with high accuracy and reproducibility.

In a preferred embodiment of the present invention, the spatial frequency range (cycle/image-width) is 0 to 512 cycle/image-width when the image length is 1024 pixels.

The present invention also includes an application step of applying a skin condition improving agent and/or a cosmetic means to a subject;
a measuring step of measuring the degree of lively face estimated by the above-mentioned estimation method, which is acquired before the application step, and the degree of change in the degree of lively face after the application step;
a determining step of determining that the skin condition improving agent and/or cosmetic procedure is more suitable for the subject when the increase rate of the degree of facial vitality is higher based on the results of the measuring step;
The present invention relates to a method for selecting a skin condition improving agent and/or a cosmetic treatment, the method being characterized by having the following:

The present invention is also a method for screening skin condition improving agents based on the correlation between the intensity of spatial frequencies obtained from facial image data and the degree of facial vitality.

In a preferred embodiment of the method for screening a skin condition improving agent of the present invention, the method includes the steps of: applying a substance to the skin;
a measuring step of measuring the degree of lively face estimated by the above-mentioned estimation method, which is acquired before the application step, and the degree of change in the degree of lively face after the application step;
a determining step of determining that a skin condition improving agent having a higher rate of increase in the degree of facial vitality is a superior skin condition improving agent based on the results of the measuring step;
The present invention is characterized by having the following.

Further, the present invention for solving the above-mentioned problem provides a spatial frequency calculation means for calculating the intensity of a spatial frequency obtained by using data obtained from any channel of RGB color channel data of color face image data alone or in combination;
an evaluation means for evaluating the degree of liveliness of the face of the subject by applying the calculated spatial frequency intensity to a formula or model that indicates a correlation between the intensity of spatial frequencies obtained from face image data and a lively face, the formula or model being prepared in advance;
The present invention relates to an estimation device comprising:
According to the estimation device of the present invention, it is possible to estimate the degree of facial liveliness of a subject from facial image data of the subject with high accuracy and good reproducibility.

Further, the present invention for solving the above problem provides a computer comprising:
a spatial frequency calculation means for calculating the intensity of a spatial frequency obtained by using data obtained from any channel alone or in combination among data of RGB color channels of the color face image data;
an evaluation means for evaluating the degree of liveliness of the face of the subject by applying the calculated spatial frequency intensity to a formula or model that indicates a correlation between the intensity of spatial frequencies obtained from face image data and a lively face, the formula or model being prepared in advance;
This relates to an estimation program for functioning as a
According to the estimation program of the present invention, a computer can estimate the degree of facial liveliness of a subject from facial image data of the subject with high accuracy and good reproducibility.

The present invention provides a new technology that estimates the degree of "liveliness of a face" with high accuracy and reproducibility.

FIG. 13 is a diagram showing the relationship between the average evaluation value (hereinafter also referred to as the score value) of "lively-looking face" and age. FIG. 1 is a diagram showing a spectral pattern resulting from FFT processing of a 1024×1024 pixel color channel divided image. FIG. 13 is a diagram showing the correlation coefficient between the score value and the power value of each spatial frequency. FIG. 13 is a diagram showing the relationship between the estimated value of the multiple regression analysis result and the score value of a face that appears lively. R4 is an image in which the power value of c/iw has been removed (0x processing) or enhanced (5x processing). R6 is an image in which the power value of c/iw has been removed (0x processing) or enhanced (5x processing). R10 is an image in which the power value of c/iw has been removed (0x processing) or enhanced (5x processing). R11 is an image in which the power value of c/iw has been removed (0x processing) or enhanced (5x processing). R25 is an image in which the power value of c/iw has been removed (0x processing) or enhanced (5x processing). This is an image in which the power values of G19 c/iw, G22 c/iw, and G28 c/iw have been removed (0x processing) or enhanced (5x processing). This is an image in which the power values of B3 c/iw and B15 c/iw have been removed (0x processing) or enhanced (5x processing). This is an image in which the power values of R62 c/iw and G75 c/iw have been removed (0x processing) or enhanced (5x processing). This is an image in which the power values of B79 c/iw and B127 c/iw have been removed (0x processing) or enhanced (5x processing). This is an image in which the power values of R74 c/iw and G57 c/iw have been removed (0x processing) or enhanced (5x processing). This is an image in which the power values of B41 c/iw, B44 c/iw, and B62 c/iw have been removed (0x processing) or enhanced (5x processing). FIG. 2 is a flow chart showing a specific embodiment of the estimation method of the present invention. FIG. 2 is a hardware block diagram of the estimation device of the present invention. 13 is a graph showing the evaluation results of "degree of facial liveliness." FIG. 13 is a representative image of the images used in the evaluation of "degree of facial liveliness." 1 is a graph showing the evaluation results of "cheek dullness." 1 is a graph showing the evaluation results of "complexion." 13 is a graph showing the evaluation results of "firmness (elasticity)." 1 is a graph showing the evaluation results of "yellowish dullness." 1 is a graph showing the evaluation results of "stiffness (roughness)."

<Method for estimating the degree of facial vivacity>
The present invention involves estimating the degree of facial liveliness of a subject from facial image data of the subject based on a correlation between the strength of spatial frequencies derived from the facial image data and the degree of facial liveliness.

The term "lively face" in the present invention refers to a face that is evaluated as lively based on the objective impression of a third party.
Specifically, a "lively face" in the present invention refers to a person who "appears lively" even with a straight face, and refers to a face that is evaluated as lively based on the objective impression of a third party, even if the person is not smiling or showing other facial expressions.

As described in the present invention, the "estimate of a lively face" includes, as shown in the examples described below, preferably 2 or more, more preferably 3 or more, even more preferably 5 or more, and particularly preferably all of the following selected from the brightness of the cheeks, dullness under the eyes, dullness of the cheeks, plumpness of the cheeks, complexion, firmness, yellowish dullness of the entire face, roughness of the skin, and gloss of the skin.

"Facial image data" can be obtained in accordance with established methods. One method for obtaining "facial image data" is, for example, to capture a frontal image of a face using a digital camera or the like, and to import the captured image information into a computer as digital information.

The subject's facial image data used in the present invention can be facial image data of the entire face or parts of the face.

When using face image data of the entire face, it is more preferable to use face data excluding the hair portion, which allows the degree of liveliness of the subject's face to be estimated with high accuracy and reproducibility from the face image data of the subject.
In this case, taking into consideration the sampling theorem for calculating the resolution and spatial frequency intensity required to estimate a lively face or the degree of livelyness of a face, it is sufficient for the size of the facial image to be captured in an area of approximately 30 cm x 30 cm, and it is preferable for the image to be 512 x 512 pixels or more.

Facial image data of facial parts includes facial image data of the cheeks, under the eyes, lips, eyes, nose, forehead, cheeks, neck, etc.
In particular, face image data of the cheeks and/or under the eyes can be preferably used.

As will be shown in the examples below, the intensity of spatial frequencies obtained from face image data of the cheeks and/or under the eyes is highly correlated with a lively face.
Therefore, by using facial image data of the cheeks and/or under the eyes, it is possible to estimate the degree of liveliness of a subject's face from the facial image data of the subject with high accuracy and reproducibility.

When using images of the cheeks and/or under the eyes, it is preferable to use images of at least an area of approximately 2 cm x 2 cm in these areas. By acquiring information regarding spatial frequency obtained from facial image data of the cheeks and/or under the eyes, it is possible to estimate the degree of liveliness of the subject's face from the facial image data of the subject with high accuracy and reproducibility. In these cases, the number of pixels in the facial image data is preferably 256 pixels or more, and more preferably 512 pixels or more.

The present invention preferably uses color channel data of either R (red), G (green), or B (blue) of a facial image to make the estimation.

In this specification, "any of the R (red), G (green), or B (blue) color channel data of a facial image" refers to color channel data that retains a specific hue among facial image data obtained by dividing a color facial image into RGB color channels.

One method for obtaining color channel data of a facial image, R (red), G (green), or B (blue), is to provide a color facial image to an image processing system and perform color channel division.

By using the color channel data of the R (red), G (green), or B (blue) of a facial image, it is possible to estimate the degree of facial liveliness of a subject from the facial image data of the subject with high accuracy and reproducibility.

A preferred embodiment of the estimation method of the present invention involves effectively utilizing information regarding the brightness of the cheeks, dullness under the eyes, dullness of the cheeks, plumpness of the cheeks, roughness of the skin, and skin luster to estimate the above-mentioned "lively face."
In the present invention, when estimating the above-mentioned "lively face" by effectively utilizing information regarding the brightness of the cheeks, dullness under the eyes, dullness of the cheeks, plumpness of the cheeks, roughness of the skin, and skin gloss, it is preferable to use the R (red) color channel data out of the R (red), G (green), and B (blue) color channel data of the facial image.

A preferred embodiment of the estimation method of the present invention includes effectively utilizing information on firmness, skin roughness, and skin gloss to estimate the above-mentioned "lively face."
In the present invention, when estimating the above-mentioned "lively face" by effectively utilizing information regarding firmness, skin roughness, and skin gloss, it is preferable to use the G (green) color channel data of the R (red), G (green), and B (blue) color channel data of the facial image.

In addition, a preferred embodiment of the estimation method of the present invention includes effectively utilizing information regarding the yellowish dullness of the entire face, the roughness of the skin, and the luster of the skin to estimate the above-mentioned "lively face."
In the present invention, when estimating the above-mentioned "lively face" by effectively utilizing information regarding the yellowish dullness of the entire face, the roughness of the skin, and the luster of the skin, it is preferable to use the B (blue) color channel data out of the R (red), G (green), and B (blue) color channel data of the facial image.

In the present invention, the spatial frequency is preferably a combination of spatial frequencies obtained from preferably two or more, and more preferably three, types of color channel data of RGB color channel data of a face image.
By using the spatial frequency intensity obtained from multiple color channel data out of the RGB color channel data of the face image, it is possible to estimate the degree of liveliness of the subject's face with high accuracy and good reproducibility.

"Spatial frequency" is a property of a structure that has a spatial periodicity, and refers to the number of repetitions of the structure contained in a unit length. Here, "spatial frequency" refers to the period per meter in the International System of Units, but in the field of image processing, the number of lines per millimeter is used.
Alternatively, the number of periods per image length (cycle/image-width) can be used as a unit of spatial frequency.
The maximum number of cycles contained in an image is half the number of pixels (pixels, dots) contained in the length of the image. That is, if the image length is 512 pixels, the spatial frequency band is 0 to 256 cycles/image-width, and if the image length is 1024 pixels, the spatial frequency band is 0 to 512 cycles/image-width.

In this specification, when referring to the spatial frequency domain of any one of the color channel data of R (red), G (green), and B (blue) of a face image, it is expressed as "hue spatial frequency domain."
For example, if the spatial frequency range obtained from the R (red) color channel data of a face image among the R (red), G (green), and B (blue) color channel data is 4 c/iw, it is expressed as "R 4 c/iw" (see Table 1, Figure 5).

"Spatial frequency intensity" refers to the amplitude value or power (power logarithm) at each spatial frequency. The spatial frequency intensity can be calculated according to common methods.

Specifically, the facial image data obtained by the above-mentioned method is divided into RGB color channels, and then FFT (FastFouRieRTRansfoRm: fast Fourier transform) is performed using commercially available image analysis software (for example, the Fourier phase analysis program made by ImageSense or PoplmaGinG made by Digital Being Kids) to calculate the amplitude value or power (spatial frequency intensity).

Figure 2 shows an example of spatial frequency analysis, which is a spectral pattern resulting from FFT processing of a 1024 x 1024 pixel gray image (or an image divided into color channels in the case of color face image data).

By FFT processing, it is possible to calculate the intensity (amplitude value or power value) for each spatial frequency (unit: cycle/image-width) from 0 to 1024/2.
The calculated intensity (amplitude value or power value) of each spatial frequency reveals the spatial frequency characteristics of the face image.

The major advantage of the present invention is that it is possible to easily obtain the numerical data (spatial frequency intensity (power spectrum)) required to estimate the degree of facial liveliness from the facial image data of a subject. In the present invention, quantitative calculations can be easily performed using commercially available software.

In the present invention, the "estimate of the degree of face liveliness" may be a quantitative estimate, or a relative and qualitative estimate based on a pre-set "face liveliness index."

Specifically, a preferred embodiment of a quantitative estimation method is one in which an equation or model is created in advance that shows the correlation between the intensity of spatial frequencies obtained from facial image data and the degree of facial liveliness, and the intensity of spatial frequencies obtained from any one of the color channel data of R (red), G (green), or B (blue) acquired from the subject is compared with the equation or model to estimate the degree of facial liveliness of the subject. By adopting such an embodiment, the estimation accuracy can be improved.

Here, the correlation between the intensity of spatial frequencies obtained from face image data and the degree of facial liveliness can be determined in advance by multivariate analysis such as regression analysis. As the multivariate analysis, one that can utilize the relationship between the objective variable and the explanatory variables is preferable, and preferred examples include discriminant analysis and regression analysis (MLR, PLS, PCR, logistic).

Among these, multiple regression analysis (MLR) and nonlinear regression analysis (PLS: PartialLeastSquaRes) are particularly preferred. For example, a multiple regression equation can be obtained by performing multiple regression analysis with the intensity of spatial frequency as the explanatory variable and a lively face as the objective variable. Also, a prediction equation (prediction model) can be obtained by performing PLS in a similar manner.

Below, we explain how to prepare an equation or model that shows the correlation between the intensity of spatial frequencies obtained from face image data and the degree of facial liveliness.

As a method for obtaining an accurate formula or model, multivariate analyses such as principal component analysis, factor analysis, quantification theory type 1, quantification theory type 2, quantification theory type 3, multidimensional scaling, supervised clustering, neural networks, and ensemble learning methods can be used as appropriate. Among these, neural networks, discriminant analysis, and quantification theory type 1 are preferred. These multivariate analyses can be performed using free software or commercially available software.

In order to obtain a highly accurate equation or model, a storage unit stores, as input information, a correlation between the intensity of spatial frequencies obtained from face image data and the degree of liveliness of a face evaluated in advance by visual inspection;
A learning unit that generates a trained model in which variables for estimating a degree of a lively face of a subject from facial image data of the subject are machine-learned using the correlation as training data;
It is also possible to use a learning device equipped with the above.

In order to create an equation or model showing the correlation between the intensity of spatial frequencies obtained from face image data and the degree of facial liveliness, it is preferable to create a database (DB) that associates the correlation between the intensity of spatial frequencies obtained from face image data and the degree of facial liveliness, where the number of people in the DB is preferably 10 or more, more preferably 20 or more, more preferably 50 or more, more preferably 100 or more, and even more preferably 200 or more.
If the DB has a matrix structure, for example, the intensities of spatial frequencies obtained from face image data can be input in the rows, and the degree of face liveliness can be input in the columns.

This DB may be updated by estimating the correlation between the intensity of spatial frequencies obtained from newly acquired facial image data of the subject and the degree of facial liveliness, and then adding the estimated value. If necessary, the above-mentioned multivariate analysis may be performed on the updated DB to update the formula or model.

Below, we will explain how to obtain the information required to create a formula or model, or to create the above DB.

(1) Evaluation of a lively face First, in order to obtain an equation or model showing the correlation between the intensity of spatial frequency obtained from face image data and the degree of lively face, an evaluator visually evaluates the lively faces of multiple people. The number of people to be evaluated is at least 50 or more, preferably 100 or more, and more preferably 500 or more. It is also preferable that the number of people is distributed equally for each age group. Since there is often a gap between the evaluation of a lively face by the person himself and the evaluation by a third party, it is preferable that the evaluator is a third party. The selection of the evaluator may be from the viewpoint of selecting an evaluator suitable to represent a third party, and it is important to consider various issues such as the individual differences of the evaluators, gender differences, lively face differences, preferences, and even reproducibility.
For example, evaluators with experience and expertise in the beauty industry are preferred. The number of evaluators is preferably more than one, specifically, about 5 to 10. The evaluation results by the multiple evaluators are preferably processed statistically.

In addition, the average value, median value, etc., excluding outliers, can be calculated to obtain an objective degree of lively facial expression for each person.
It is desirable to evaluate the degree of facial liveliness using a facial image, as shown in the examples described later, because this makes it possible to eliminate the influence of facial expressions, hairstyles, backgrounds, skin color, skin problems, etc., which may reduce the accuracy of the evaluation.

(2) Calculation of spatial frequency intensity Next, the spatial frequency intensity is calculated from the face image data of the plurality of people. The calculation of the spatial frequency intensity is as described above. It is preferable that the face image used here is the same as the face image used in the evaluation of the lively face described above. This makes it possible to obtain a more accurate correlation between the spatial frequency intensity and the lively face level.

Below, we will explain in detail a preferred embodiment of the spatial frequency used in the estimation method of the present invention.

The estimation method of the present invention preferably uses the intensity of spatial frequencies at spatial frequencies (cycles/image-width) that are preferably 1/5 or less of the image length (pixels), more preferably 1/10, and even more preferably 1/15.

Unless otherwise specified, the following description refers to the preferred form when using facial image data with an image length of 1024 pixels.

In addition, the estimation method of the present invention preferably uses the intensities of spatial frequencies of 2 or more, more preferably 3 or more, more preferably 5 or more, more preferably 8 or more, even more preferably 10 or more, even more preferably 15 or more, and particularly preferably all spatial frequencies.

In the estimation method of the present invention, it is preferable to use the spatial frequency intensity of one or more, preferably two or more, more preferably three or more, more preferably five or more, more preferably eight or more, even more preferably ten or more, even more preferably fifteen or more, and particularly preferably all spatial frequencies selected from R4 c/iw, R6 c/iw, R10 c/iw, R11 c/iw, R25 c/iw, G19 c/iw, G22 c/iw, G28 c/iw, B3 c/iw, B15 c/iw, R62 c/iw, G75 c/iw, B79 c/iw, B127 c/iw, R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw.

In addition, the estimation method of the present invention preferably uses the spatial frequency intensity of one or more, preferably two or more, more preferably three or more, more preferably five or more, more preferably eight or more, even more preferably ten or more, and particularly preferably all spatial frequencies selected from R4 c/iw, R11 c/iw, G22 c/iw, R62 c/iw, G75 c/iw, B79 c/iw, B127 c/iw, R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw.

In addition, the estimation method of the present invention preferably uses the spatial frequency intensity at one or more, preferably two or more, more preferably three or more, and particularly preferably all spatial frequencies selected from R62 c/iw, G75 c/iw, G57 c/iw, B44 c/iw, and B62 c/iw.

In a particularly preferred embodiment of the present invention, the higher the intensity of the R4 c/iw spatial frequency, the brighter the cheeks are judged to be, and the face is estimated to be lively based on this judgment.

In a particularly preferred embodiment of the present invention, the higher the intensity of the R6 c/iw spatial frequency, the better the brightness under the eyes is determined to be, and the face is estimated to be lively based on this determination.

In a particularly preferred embodiment of the present invention, the lower the intensity of the spatial frequency R10 c/iw, the less dull the cheeks are judged to be, and based on this judgment, the face is estimated to be lively.

In a particularly preferred embodiment of the present invention, the higher the intensity of the spatial frequency R11 c/iw, the plumper the cheeks are judged to be, and the face is estimated to be lively based on this judgment.

In a particularly preferred embodiment of the present invention, the lower the intensity of the R25 c/iw spatial frequency, the better the complexion is judged to be, and the face is estimated to be lively based on this judgment.

In a particularly preferred embodiment of the present invention, the lower the intensity of one or more, preferably two or more, and more preferably three or more spatial frequencies selected from G19 c/iw, G22 c/iw, and G28 c/iw, the greater the firmness (elasticity) of the cheeks (whole face), and the face is estimated to be lively based on this determination.

In a particularly preferred embodiment of the present invention, the lower the intensity of one or two spatial frequencies selected from B3 c/iw and B15 c/iw, the less yellowish dullness there is overall on the face, and the face is estimated to be lively based on this determination.

In a particularly preferred embodiment of the present invention, the lower the intensity of one or more, preferably two or more, more preferably three or more, and even more preferably all spatial frequencies selected from R62 c/iw, G75 c/iw, B79 c/iw, and B127 c/iw, the less stiff (stiff) the face is judged to be, and the face is estimated to be lively based on this judgment.

In a particularly preferred embodiment of the present invention, the lower the intensity of one or more, preferably two or more, more preferably three or more, and even more preferably all spatial frequencies selected from R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw, the better the skin luster of the entire face is judged to be, and the face is estimated to be lively based on this judgment.

The specific flow of the estimation method of the present invention is explained below.

In addition, as a specific embodiment of the estimation method of the present invention,
A facial image acquisition step S1 of acquiring facial image data from a subject;
a channel division step S2 for acquiring R (red), G (green), and B (blue) color channel data from the face image data acquired in the face image acquisition step;
a spatial frequency calculation step S3 of calculating the intensity of a spatial frequency from at least one of the color channel data of RGB;
an evaluation step S4 of applying the calculated spatial frequency intensity to a formula or model prepared in advance that indicates a correlation between the intensity of spatial frequencies obtained from face image data and the degree of lively face to evaluate the degree of lively face of the subject;
a display step S5 for displaying the evaluation result in the evaluation step S4;
(See FIG. 16 ).

Here, in the present invention, it is not necessary to carry out all of the above steps S1 to S5. For example, a spatial frequency calculation step S3 of calculating the intensity of a spatial frequency from at least one of the RGB color channel data of the face image,
The method may include an evaluation step S4 in which the calculated spatial frequency intensity is applied to a formula or model prepared in advance that indicates the correlation between the intensity of spatial frequency obtained from facial image data and the degree of facial liveliness, thereby evaluating the degree of facial liveliness of the subject.

In the present invention, it is preferable that the method includes updating a DB (database) based on the results of the evaluation in the evaluation step S4 and/or storing the results of the evaluation in the evaluation step S4 in a DB (database).
In particular, it is preferable that the above-mentioned updating and accumulation is fed back and evaluated in the evaluation step S4.

The degree of facial vitality estimated by this invention can be used as an indicator when selecting cosmetics.

Specifically, the present invention relates to a method for treating a skin condition by applying a skin condition improving agent and/or a cosmetic means to a subject;
a measuring step of measuring the degree of lively face estimated by the above-mentioned estimation method, which is acquired before the application step, and the degree of change in the degree of lively face after the application step;
A determining step of determining that the substance and/or cosmetic means exhibits a skin condition improving effect suitable for the subject as the increase rate of the degree of lively face is higher based on the results of the measuring step;
The present invention also relates to a method for selecting a skin condition improving agent and/or a cosmetic treatment, the method being characterized by having the following:

Skin condition improving agents include commercially available compounds (including peptides), known compounds (including peptides), compounds obtained by combinatorial chemistry techniques, natural ingredients derived from plants or marine organisms, and animal tissue extracts.

The extract derived from an animal or plant is a general term not only for the extract itself derived from an animal or plant, but also for a fraction of the extract, a purified fraction, and a product obtained by removing the solvent from the extract, fraction, or purified product.
Examples of plant-derived extracts include extracts using wild-growing or cultivated plants, extracts using plants sold as raw materials for herbal medicines, and commercially available extracts.

Examples of beauty treatments include face packs and massage techniques.

The results of the degree of facial vitality determined by the present invention can also be a useful indicator for counseling regarding skin care and makeup methods.

Here, the above explanation can be used as a preferred embodiment of the method for selecting a skin condition improving agent and/or cosmetic treatment of the present invention.

<Screening method for skin condition improving agents>
The present invention also provides a method for screening skin condition improving agents based on the correlation between the intensity of spatial frequencies obtained from face image data and the degree of facial vitality.

The present invention relates to
applying the substance to the skin;
a measuring step of measuring the degree of lively face estimated by the above-mentioned estimation method, which is acquired before the application step, and the degree of change in the degree of lively face after the application step;
a determining step of determining that the higher the rate of increase in the degree of facial vitality based on the results of the measuring step, the better the skin condition improving agent is;
It is also possible to adopt a form having the above structure.

The present invention also relates to a method for treating a disease comprising the steps of: applying two or more substances to two or more subjects;
a measuring step of measuring, after the application step, a degree of change in the degree of lively face obtained before the application step and a degree of lively face obtained after the application step;
a determining step of determining that a skin condition improving agent having a higher rate of increase in the degree of facial vitality is a superior skin condition improving agent based on the measurement results of the measuring step;
It is also possible to adopt a form having the above structure.

In the screening method of the present invention, the agents for suppressing and/or improving spots to be screened are not particularly limited, and examples thereof include commercially available compounds (including peptides), known compounds (including peptides), compounds obtained by combinatorial chemistry techniques, natural components derived from plants or marine organisms, and animal tissue extracts.

The extract derived from an animal or plant is a general term not only for the extract itself derived from an animal or plant, but also for a fraction of the extract, a purified fraction, and a product obtained by removing the solvent from the extract, fraction, or purified product.
Examples of plant-derived extracts include extracts using wild-growing or cultivated plants, extracts using plants sold as raw materials for herbal medicines, and commercially available extracts.

In the present invention, it is particularly preferable to screen agents containing plant-derived extracts.

In the present invention, the types of candidate substances to be applied may be two or more, preferably three or more.
According to the present invention, when a plurality of candidate substances are applied, it is possible to screen for a skin condition improving agent that exerts a synergistic skin condition improving effect.

Furthermore, the screening method of the present invention can also be used to screen products in the form of cosmetics.
Here, the cosmetic product is preferably one that is applied to the skin. Preferred examples of the cosmetic product include wash, lotion, and emulsion.

In the present invention, the period of the application step is preferably 2 weeks or more, more preferably 3 weeks or more, further preferably 6 weeks or more, and particularly preferably 10 weeks or more.
By carrying out the application step for the above-mentioned period or longer, skin condition improving agents can be screened more accurately.

Here, the above explanation can be used as a preferred embodiment of the screening method of the present invention.

<Device for estimating the degree of facial liveliness>
The estimation device of the present invention will be described below with reference to a hardware block diagram (FIG. 17) of the estimation device of the present invention.

The estimation device 1 of the present invention comprises a facial image acquisition means 11 for acquiring facial image data of a subject, a channel division means 12 for dividing the acquired facial image data into specific color channels, a spatial frequency calculation means 13 for calculating the intensity of spatial frequencies obtained from any channel of the RGB color channel data of the subject's color facial image data, either alone or in combination, a memory unit 14 for storing correlation data indicating the correlation between the intensity of spatial frequencies obtained from the facial image data and the degree of facial liveliness, an evaluation means 15, and a display means 16 for displaying the evaluation results of the evaluation means 15.

The facial image acquisition means 11 is a means for acquiring a facial image obtained by a device for inputting facial image data, such as a digital camera, a digital microscope, a digital video, or a scanner.

The channel splitting means 12 is a means for splitting the color face image data acquired by the face image acquisition means 11 into RGB color channels and generating grayscale image data (RGB color channel data of the face image) that retains one of the RGB hues in order to provide the face image data to the spatial frequency calculation means 13 described below.

The spatial frequency calculation means 13 performs processes such as calculating the intensity of spatial frequencies by FFT processing, and calculating an equation or model (correlation data) that shows the relationship between a lively face (a lively face or a lively face) and the intensity of spatial frequencies.

The storage unit 14 stores programs necessary for the functioning of the face liveliness estimation device of the present invention, and various equations or models (correlation data) necessary for estimating the face liveliness. The storage unit 14 may further store a database necessary for creating the equations or models.

The evaluation means 15 is a means for evaluating the degree of liveliness of the subject's face based on the intensity of the spatial frequency calculated by the spatial frequency calculation means 13 and the correlation data stored in the memory unit 14.

The display means 16 is a means for instructing an arbitrary display device to display the evaluation result by the evaluation means 15 .
The display device may be anything capable of displaying the estimation results of the degree of face liveliness, and examples of such a display device include a CRT (Cathode Ray TuBu) or liquid crystal display, an audio output device such as a speaker, or a printer.

In the present invention, the above configuration is not necessarily required. For example, a spatial frequency calculation means 13 for calculating the intensity of a spatial frequency obtained from at least one of the RGB color channel data of a face image,
an evaluation means for evaluating the degree of liveliness of the face of the subject by applying the calculated spatial frequency intensity to a formula or model that indicates a correlation between the intensity of spatial frequencies obtained from face image data and a lively face, the formula or model being prepared in advance;
The present invention may be in the form of an estimation device comprising:

The present invention may also take the form of an estimation device that includes estimating the degree of facial liveliness of a subject from facial image data of the subject based on the correlation between the intensity of spatial frequencies obtained from the facial image data and the degree of facial liveliness.

Here, the above description can be used as a preferred embodiment of the estimation device of the present invention.

<Program for estimating the degree of facial liveliness>
The estimation program of the present invention causes a computer, other device, machine, etc. to function as a facial image acquisition means 11 that acquires facial image data of a subject, a channel division means 12 that divides the acquired facial image data into specific color channels, a spatial frequency calculation means 13 that calculates the intensity of spatial frequencies obtained from any channel of the RGB color channel data of the subject's color facial image data, either alone or in combination, a memory unit 14 that stores correlation data indicating the correlation between the intensity of spatial frequencies obtained from the facial image data and the degree of liveliness of the face, an evaluation means 15, and a display means 16 that displays the evaluation results of the evaluation means 15 (see Figure 17).

In the present invention, it is not necessary for the program to operate all of the above components. For example, a computer may be configured to have a spatial frequency calculation means 13 for calculating the intensity of a spatial frequency obtained by using data obtained from any channel of the RGB color channel data of the color face image data alone or in combination.
an evaluation means for evaluating the degree of liveliness of the face of the subject by applying the calculated spatial frequency intensity to a formula or model that indicates a correlation between the intensity of spatial frequencies obtained from face image data and a lively face, the formula or model being prepared in advance;
The present invention may be in the form of an estimation program that functions as a

The present invention also provides a computer comprising:
The present invention may take the form of an estimation program that functions to estimate the degree of livelyness of a subject's face from the facial image data of the subject, based on the correlation between the intensity of spatial frequencies obtained from the facial image data and the degree of livelyness of the face.

The present invention may also be implemented as a program recorded on a recording medium that can be read by a computer or the like.

Here, the above explanation can be used as a preferred embodiment of the estimation program of the present invention.

The test results on which the present invention is based are shown below.

<1> Evaluation of lively faces (1) Preparation of images to be evaluated First, facial images of 272 women aged 18 to 82 years (average age 49.1 years, SD = 16.91) were photographed. Based on the photographed facial images, 1024 x 1024 pixel image data for testing was created. Here, in creating the images to be used for testing, hair and ears were removed from the original images, the left and right pupils were fixed to fixed positions in the image, and the size and position of the face were standardized.

(2) Impression Evaluation Twenty healthy women (5 in their 20s, 5 in their 30s, 5 in their 40s, and 5 in their 50s, average age 39.6 years, SD=11.3) were used to evaluate the "lively face."
The evaluators rated the face images prepared in (1) on whether or not the people looked "lively" using the following 7-point scale.
In order to avoid bias in the evaluators' responses, the order in which the facial images were presented was changed for each evaluator and used in the test. The average evaluation value (hereinafter referred to as the score value) of "lively-looking face" for each facial image was calculated.

Score 1: Not very vivid Score 2: Not very vivid Score 3: Not quite vivid Score 4: Undecided Score 5: Somewhat vivid Score 6: Quite vivid Score 7: Very vivid

(3) Relationship between score value and age The relationship between score value and age was plotted on the vertical axis and age on the horizontal axis (Figure 1).
As shown in FIG. 1, the score values did not decrease with age.
In other words, there was no correlation between the score and age.

(4) Relationship between score value and intensity of each spatial frequency (4-1) Spatial frequency analysis of face image First, the color face image used in the test in (2) was divided into RGB color channels using an image processing system to obtain color channel data of R (red), G (green), and B (blue) (corresponding to face image data in this specification). Of R (red), G (green), and B (blue), color channel data that retained a specific hue was used for spatial frequency analysis software developed by TheMathWorks, Inc. using MATLAB (registered trademark), and spatial frequency analysis was performed.

By performing spatial frequency analysis using MATLAB (registered trademark), power values (spatial frequency intensity) of 1 to 512 cycles/image-width (hereinafter, c/iw) were obtained for the color channel data of R (red), G (green), or B (blue) of the face images used in the test in (2).

(4-2) Deriving the correlation coefficient between the score value and the power value of each spatial frequency The correlation coefficient between the score value and the power value of each spatial frequency was derived for each color channel data of R (red), G (green), or B (blue). The results are shown in Figure 3.
As shown in FIG. 3, a significant power value correlated with the score value was confirmed.

(4-3) Multiple Regression Analysis Multiple regression analysis was performed on the score values and the power values of each spatial frequency to confirm characteristic power values.
Here, the multiple regression analysis was performed by a stepwise method using the score value as the objective variable and the power value of each spatial frequency as the explanatory variable.
The results are shown in Table 1. In Table 1, items are selected for which the sign (positive or negative) of the estimated value of the multiple regression analysis result (hereinafter simply referred to as the estimated value) matches the sign (positive or negative) of the correlation coefficient shown in Figure 3.

Hereinafter, in this embodiment, when referring to the spatial frequency range of data obtained from color channel data that retains a specific hue among any of the color channel data of R (red), G (green), and B (blue), it will be expressed as the "hue spatial frequency range of the divided color channel."
For example, when the spatial frequency range of data obtained from the R (red) color channel data among the R (red), G (green), and B (blue) color channel data is 4 c/iw, it is expressed as "R 4 c/iw" (see Table 1, Figure 5).

R-squared: 0.444658
Degree of freedom adjusted R-squared: 0.402787
Analysis of variance: F value < 0.0001

Estimation formula: =-2.528026029+1.7914917386xR4+0.9038084107xR6-1.043600467xR10+1.6881883003xR11-1.052159449xR25-2.241496576xR62+1.95489 54779xR74-0.647420451xG19-1.634556714xG2 2-1.327723116xG28+3.4750220608xG57-2.780306003xG75-0.670958138xB3-0.960381395xB15+1.8566437495xB41+2.09147472xB44+3.70297 02782xB62-1.876045866xB79-1.65718134xB127

For each face image data acquired in (1), the score value is set on the vertical axis and the estimated value by the multiple regression equation (hereinafter simply referred to as the estimated value) is set on the horizontal axis, and the relationship between the score value and the estimated value is illustrated (Figure 4).
As shown in FIG. 4, it was found that the higher the score value, the higher the estimated value.
That is, it was found that the multiple regression equation derived in this embodiment shows the correlation between the score value (degree of lively face) and the estimated value.

<2> Verification of the estimated value of the degree of facial liveliness We verified the results derived in <1>.

An investigation was carried out to determine what features in the facial impression each power value at the spatial frequencies shown in Table 1 reflects.
For the above verification, an image was created in which the power value at a specific spatial frequency was removed or enhanced, and the image was subjected to evaluation by experts who are experts in technologies related to skin conditions.

(1) Verify what features of face images are reflected by the power value of the spatial frequency range R4 c/iw

As shown in Table 1, the correlation coefficient between the power value of R4 c/iw and a lively face is 1.791. The higher the power value of R4 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power value of R4 c/iw was removed (0x processing) or enhanced (5x processing). The results are shown in FIG.
As shown in FIG. 5, it was found that the higher the power value of R4 c/iw, the more lively the face becomes.

As shown in FIG. 5, it was found that the cheek area is dark in the image in which the power value has been removed (processed with a factor of 0), and is bright in the image in which the power value has been removed (processed with a factor of 5).
In other words, it was found that the power value of R4 c/iw is highly related to the cheek area.
And the power value of R4 c/iw changed with respect to the brightness of the cheeks.

(2) Verify what features of face images are reflected by the power value of the spatial frequency range R6 c/iw

As shown in Table 1, the correlation coefficient between the power value of R6 c/iw and a lively face is 0.0965. The higher the power value of R6 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power value of R6 c/iw was removed (0x processing) or enhanced (5x processing). The results are shown in FIG.
As shown in FIG. 6, it was found that the higher the power value of R6 c/iw, the more lively the face becomes.

As shown in FIG. 6, it was found that the area under the eyes is dark in the image where the power value has been removed (processed at a factor of 0), whereas the area under the eyes is bright in the image where the power value has been removed (processed at a factor of 5).
That is, it was found that the power value of R6 c/iw is highly relevant under the eyes.
And, it was found that the power value of R6 c/iw is related to brightness under the eyes.

(3) Verify what features of face images are reflected by the power value of the spatial frequency range R10 c/iw

As shown in Table 1, the correlation coefficient between the power value of R10 c/iw and a lively face is -0.0118. The lower the power value of R10 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power value of R10 c/iw was removed (0x processing) or enhanced (5x processing). The results are shown in FIG.
As shown in FIG. 7, it was found that the lower the power value of R6 c/iw, the more lively the face becomes.

As shown in FIG. 7, it was found that the cheeks were bright in the image in which the power value was removed (processed at a factor of 0), and the cheeks were dull in the image in which the power value was removed (processed at a factor of 5).
That is, it was found that the power value of R10 c/iw is highly related to the cheek.
It was also found that the power value of R10 c/iw is related to dullness of the cheeks.

(4) Verify what features of face images are reflected by the power value of the spatial frequency range R11 c/iw

As shown in Table 1, the correlation coefficient between the power value of R11 c/iw and a lively face is 0.1487. The higher the power value of R11 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power value of R11 c/iw was removed (0x processing) or enhanced (5x processing). The results are shown in FIG.
As shown in FIG. 8, it was found that the higher the power value of R11 c/iw, the more lively the face becomes.

As shown in FIG. 8, the image with the power value removed (0x processing) did not give the cheeks a plump appearance, whereas the image with the power value removed (5x processing) gave the cheeks a plump appearance.
That is, it was found that the power value of R11 c/iw is highly related to the cheek.
It was also found that the power value of R11 c/iw is related to the plumpness of the cheeks.

(5) Verify what features of face images are reflected by the power value of the spatial frequency range R25 c/iw

As shown in Table 1, the correlation coefficient between the power value of R25 c/iw and a lively face is -0.1887. The lower the power value of R25 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power value of R25 c/iw was removed (0x processing) or enhanced (5x processing). The results are shown in FIG.
As shown in FIG. 9, it was found that the lower the power value of R25 c/iw, the more lively the face becomes.

As shown in FIG. 9, it was found that in the image where the power value was removed (0-fold processing), the entire face looked healthy, whereas in the image where the power value was removed (5-fold processing), the entire face looked poor in color.
That is, it was found that the power value of R25 c/iw is highly relevant to the entire face.
It was also found that the power value of R25 c/iw is related to good complexion.

(6) Verification of what features in facial images are reflected by the power values of the spatial frequency ranges G19 c/iw, G22 c/iw, and G28 c/iw.

As shown in Table 1, the correlation coefficient between the power value of G19 c/iw and a lively face is -0.1325. The lower the power value of G19 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of G22 c/iw and a lively face is -0.2519. The lower the power value of G22 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of G28 c/iw and a lively face is -0.1685. The lower the power value of G28 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power values of G19 c/iw, G22 c/iw, and G28 c/iw were removed (0x processing) or enhanced (5x processing). The results are shown in FIG.
As shown in FIG. 10, it was found that the lower the power values of G19 c/iw, G22 c/iw, and G28 c/iw, the more lively the face becomes.

As shown in Figure 10, the image with the power value removed (0x processing) shows firmness (elasticity) in the cheeks (entire face), whereas the image with the power value removed (5x processing) shows a lack of firmness (elasticity) in the cheeks (entire face).
That is, it was found that the power values of G19 c/iw, G22 c/iw, and G28 c/iw are highly related to the entire face.
It was also found that the power values of G19 c/iw, G22 c/iw, and G28 c/iw relate to firmness (elasticity) of the cheeks (entire face).

(7) Verification of what features in facial images are reflected by the power values of the spatial frequency ranges B3 c/iw and B15 c/iw

As shown in Table 1, the correlation coefficient between the power value of B3 c/iw and a lively face is -0.1042. The lower the power value of B3 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of B15 c/iw and a lively face is -0.1752. The lower the power value of B15 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power values of B3 c/iw and B15 c/iw were removed (0x processing) or enhanced (5x processing). The results are shown in FIG.
As shown in FIG. 11, it was found that the lower the power values of B3 c/iw and B15 c/iw, the more lively the face becomes.

As shown in FIG. 11, in the image in which the power value was removed (0x processing), there was no yellowish dullness over the entire face, whereas in the image in which the power value was removed (5x processing), there was a yellowish dullness over the entire face.
That is, it was found that the power values of B3 c/iw and B15 c/iw are highly related to the whole face.
It was also found that the power values of B3 c/iw and B15 c/iw relate to yellowish dullness of the entire face.

(8) Verification of what features in facial images are reflected by the power values of the spatial frequency ranges R62 c/iw, G75 c/iw, B79 c/iw, and B127 c/iw.

As shown in Table 1, the correlation coefficient between the power value of R62 c/iw and a lively face is -0.0334. The lower the power value of R62 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of G75 c/iw and a lively face is -0.1973. The lower the power value of G75 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of B79 c/iw and a lively face is -0.1033. The lower the power value of B79 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of B127 c/iw and a lively face is -0.2388. The lower the power value of B127 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power values of R62 c/iw, G75 c/iw, B79 c/iw, and B127 c/iw were removed (0x processing) or enhanced (5x processing). The results are shown in Figures 12 and 13.
As shown in Figs. 12 and 13, it was found that the lower the power values of R62 c/iw, G75 c/iw, B79 c/iw, and B127 c/iw, the more lively the face became.

As shown in Figures 12 and 13, in the image where the power value was removed (0x processing), the entire face appeared smooth, whereas in the image where the power value was removed (5x processing), the entire face appeared stiff (rough).
That is, it was found that the power values of R62 c/iw, G75 c/iw, B79 c/iw, and B127 c/iw are highly related to the entire face.
It was also found that the power values of R62 c/iw, G75 c/iw, B79 c/iw, and B127 c/iw relate to stiffness.

(9) Verification of what features in facial images are reflected by the power values of the spatial frequency ranges R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw.

As shown in Table 1, the correlation coefficient between the power value of R74 c/iw and a lively face is 0.0598. The higher the power value of R74 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of G57 c/iw and a lively face is 0.0182. The higher the power value of G57 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of B41 c/iw and a lively face is 0.0922. The higher the power value of B41 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of B44 c/iw and a lively face is 0.0717. The higher the power value of B44 c/iw, the more lively the face is evaluated to be.

As shown in Table 1, the correlation coefficient between the power value of B62 c/iw and a lively face is 0.0177. The higher the power value of B62 c/iw, the more lively the face is evaluated to be.

To confirm the above evaluation, images were prepared in which the power values of R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw were removed (0x processing) or enhanced (5x processing). The results are shown in Figures 14 and 15.
As shown in Figs. 14 and 15, it was found that the higher the power values of R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw, the more lively the face became.

As shown in Figures 14 and 15, the image with the power value removed (0x processing) lacked the glossy smoothness of the skin over the entire face, whereas the image with the power value removed (5x processing) had the glossy smoothness of the skin over the entire face.
That is, it was found that the power values of R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw are highly related to the entire face.
It was also found that the power values of R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw are related to skin gloss.

(10) Results and Observations The results of the above verifications (1) to (9) are shown in Table 2.

The results of this example show that it is possible to estimate the degree of facial liveliness of a subject from the facial image data of the subject based on the correlation between the intensity of spatial frequencies obtained from the facial image data and the degree of facial liveliness.

In particular, the results of this embodiment show that the degree of facial liveliness can be estimated by using the power values at one or more spatial frequencies selected from R4 c/iw, R6 c/iw, R10 c/iw, R11 c/iw, R25 c/iw, G19 c/iw, G22 c/iw, G28 c/iw, B3 c/iw, B15 c/iw, R62 c/iw, G75 c/iw, B79 c/iw, B127 c/iw, R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw.

<3> Study on a method for selecting a skin condition improving agent and/or a cosmetic procedure, and a screening method for skin condition improving agents Next, based on the above findings, a study was conducted on a method for selecting a skin condition improving agent and/or a cosmetic procedure, and a screening method.

(1) Preparation Step (1-1) Preparation of Test Products First, the test products (wash, lotion, and emulsion) shown below were prepared.

[Wash]
A wash was made according to the formula shown below.

<Wash formula>
Fatty acid 40.0 parts by weight Glycerin 19.2 parts by weight Diglycerin 3.5 parts by weight Na carboxymethylcellulose 0.4 parts by weight Acrylic acid/dimethyldiallylammonium chloride/acrylamide copolymer liquid 2.0 parts by weight Dimethyldiallylammonium chloride/acrylamide copolymer liquid 4.5 parts by weight Potassium hydroxide 8.5 parts by weight Alkyl glycoside 1.5 parts by weight Plant extract 0.4 parts by weight Water 20 parts by weight

[Lotion]
A cosmetic preparation for external use on the skin was prepared according to the following recipe: The following ingredients were heated to 80° C., stirred and solubilized, and cooled with stirring to obtain a skin lotion.

<Lotion formulation>
1,2-Hexanediol 3 parts by weight 1,3-Butanediol 5 parts by weight Glycerin 2 parts by weight Phenoxyethanol 0.5 parts by weight Plant extracts 2 parts by weight Polyoxyethylene hydrogenated castor oil 0.1 parts by weight Ethanol 5 parts by weight

[Emulsion]
A cosmetic preparation for external use on the skin of the present invention was prepared according to the following recipe: In other words, ingredients A, B, and C were each heated to 70°C, B was neutralized with C, A was gradually added while stirring to emulsify, and the mixture was homogenized with a homogenizer, and then cooled with stirring to obtain an emulsion.

<Emulsion formula>
Isqualane 10 parts by weight Sorbitan sesquistearate 2 parts by weight Plant extracts 2.1 parts by weight Butylparaben 0.1 parts by weight 1,3-butanediol 5 parts by weight Xanthan gum 0.1 parts by weight Acrylic acid / alkyl methacrylate (C10-30) 0.4 parts by weight Methylparaben 0.1 parts by weight Water 50 parts by weight Potassium hydroxide 0.2 parts by weight Water 30 parts by weight

(1-2) Subjects and Conditions for Use of Test Product The subjects in this test were 19 women aged 35-49 years (average age 41.6±4.54 years).

Subjects applied the three test products (wash, lotion, and emulsion) in place of or in addition to their regular skin care products during the study period.
At this time, the subjects applied the test product to their face every day (twice a day, morning and evening) following the appropriate directions for use.
Regarding makeup products other than the above test products, the subjects continued to use their usual makeup products.

(2) Measurement process (spatial frequency analysis method)
(2-1) Acquisition of Images for Analysis Images of the face of the subject were taken 0 weeks (before use), 4 weeks, and 12 weeks after the start of use of the test product. Hair and ears were removed from the images, and the left and right pupils were fixed in fixed positions on the images, and 1024 x 1024 pixel images were created with the size and position of the face standardized (see FIG. 19).

(2-2) Acquisition and Selection of Power Values A power value of 1512 cycle/image width (hereinafter, c/iw) was obtained for each of R, G, and B by fast Fourier transform (FFT) using spatial frequency analysis (MATLAB (registered trademark) R2009a manufactured by TheMathWorks, Inc.).

From the obtained power values, the power values related to the "degree of liveliness of the face" were extracted as shown in the table. Here, the extracted power values are the power values at the spatial frequencies of R4 c/iw, R6 c/iw, R10 c/iw, R11 c/iw, R25 c/iw, G19 c/iw, G22 c/iw, G28 c/iw, B3 c/iw, B15 c/iw, R62 c/iw, G75 c/iw, B79 c/iw, B127 c/iw, R74 c/iw, G57 c/iw, B41 c/iw, B44 c/iw, and B62 c/iw.

(3) Evaluation Step Based on the measured values in the measurement step, the usefulness of the candidate substances to be screened as agents for suppressing and/or improving skin pigmentation symptoms was evaluated.

(3-1) Evaluation of "Degree of lively face" The extracted power value was substituted into the estimated value described above to calculate the change in "degree of lively face". The results are shown in Fig. 18. A representative example of the subject's face image is shown in Fig. 19.

As shown in the results of FIG. 18, it was found that the application of the test product made it possible to evaluate changes in the "degree of facial vivacity."
That is, it was found that the above method makes it possible to select a skin condition improving agent and/or a cosmetic procedure.
It was also found that the above method makes it possible to screen candidate substances for skin condition improving agents.

(3-2) Evaluation of each element of "Degree of lively face" Next, each element related to "degree of lively face" was evaluated based on the extracted power value. The results are shown in Figs. 20 to 24.

In FIG. 20, a significant decrease in the value of "cheek dullness" was observed between 0 weeks and 4 weeks.
Here, as described above, the correlation coefficient sign between the evaluation value (see the estimation formula given above) and the power value of R10 c/iw is "-".
Therefore, this evaluation resulted in the conclusion that "it can be estimated that the test product is a candidate substance for improving cheek dullness."

Furthermore, in FIG. 21, a significant decrease in the value of "complexion" was observed between weeks 0 and 4 and between weeks 0 and 12.
Here, as described above, the correlation coefficient sign between the evaluation value (see the estimation formula shown above) and the power value of R25 c/iw is "-".
Therefore, the results of this evaluation indicate that "the test product can be presumed to be a candidate substance for improving blood color."

Furthermore, in FIG. 22, a significant decrease in the value of "firmness (elasticity)" was observed between 0 and 4 weeks and between 0 and 12 weeks.
Here, as described above, the correlation coefficient sign between the evaluation value (see the estimation formula shown above) and the power values of G19 c/iw, G22 c/iw, and G28 c/iw is "-".
Therefore, the results of this evaluation indicate that "the test product can be estimated to be a candidate substance for improving firmness (elasticity)."

Furthermore, in FIG. 23, a significant decrease in the value of "yellowing dullness of the entire face" was observed between weeks 0 and 4 and between weeks 0 and 12.
Here, as described above, the correlation coefficient sign between the evaluation value (see the estimation formula shown above) and the power values of B3 c/iw and B15 c/iw is "-".
Therefore, from this evaluation, the result was that "it can be estimated that the test product is a candidate substance for improving yellowish dullness of the entire face."

Furthermore, in FIG. 24, a significant decrease in the value of "stiffness" was observed between weeks 0 and 4 and between weeks 0 and 12.
Here, as described above, the correlation coefficient sign between the evaluation value (see the estimation formula shown above) and the power values of R62 c/iw, G75 c/iw, B79 c/iw, and B127 c/iw is "-".
Therefore, from this evaluation, the result was that "it can be estimated that the test product is a candidate substance for improving the stiffness (stiffness)."

As shown in the results of Figures 20 to 24, it was found that by applying the three test products (wash, lotion, and emulsion), it was possible to evaluate changes in the "lively face" based on the judgment of each element related to the "lively face."

As described above, it was found that the above method makes it possible to select a skin condition improving agent and/or a cosmetic procedure.
It was also found that the above method makes it possible to screen candidate substances for skin condition improving agents.

The technology for estimating the degree of facial vitality of the present invention can be used in evaluating the effects of esthetics, skin care, makeup, health foods, cosmetic surgery, etc., or in providing advice or counseling using the results of such evaluation.

Claims (10)

estimating a degree of facial vivacity of the subject from the facial image data of the subject based on a correlation between the intensity of spatial frequencies obtained from the facial image data and the degree of facial vivacity ;
a spatial frequency calculation step of individually calculating the intensity of a spatial frequency obtained from color channel data having a specific hue among RGB color channel data of the color face image data of the face image, and combining and using the spatial frequencies obtained from two or more types of color channel data;
an evaluation step of evaluating the lively face of the subject by applying the calculated spatial frequency intensities to an equation or model that indicates a correlation between the intensities of two or more spatial frequencies obtained from color channel data that holds a specific hue and is obtained from face image data, the intensities being prepared in advance, and the lively face;
An estimation method, comprising : 2. The estimation method according to claim 1 , wherein the equation is a multiple regression equation obtained by multiple regression analysis or a prediction equation obtained by PLS. a facial image acquisition step of acquiring the facial image data from a subject;
a channel division step of acquiring R (red), G (green), and B (blue) color channel data from the face image data acquired in the face image acquisition step;
a spatial frequency calculation step of calculating the intensity of a spatial frequency from at least any one of RGB color channel data and using a combination of two or more of the spatial frequencies;
an evaluation step of evaluating the lively face of the subject by applying the calculated spatial frequency intensities to a formula or model prepared in advance that indicates a correlation between the intensities of two or more spatial frequencies obtained from face image data and the lively face level;
a display step of displaying the evaluation result in the evaluation step;
The estimation method according to claim 1 or 2, comprising: The method according to claim 1 , wherein the facial image data is facial image data of a portion of a face or an entire face. The estimation method according to any one of claims 1 to 4 , wherein the spatial frequency range (cycle/image-width) of the spatial frequency is 0 to 512 cycle/image-width when the image length is 1024 pixels. An application step of applying a skin condition improving agent and/or a cosmetic means to a subject;
a measuring step of measuring a degree of facial vivacity estimated by the estimation method according to any one of claims 1 to 5 , which is acquired before the application step, and a degree of change in the degree of facial vivacity after the application step;
a determining step of determining that the skin condition improving agent and/or cosmetic procedure is more suitable for the subject when the increase rate of the degree of facial vitality is higher based on the results of the measuring step;
A method for selecting a skin condition improving agent and/or cosmetic treatment, comprising the steps of: A method for screening a skin condition improving agent based on a correlation between an intensity of a spatial frequency obtained from face image data and a degree of facial vitality, comprising:
applying the substance to the skin;
a measuring step of measuring a degree of facial vivacity estimated by the estimation method according to any one of claims 1 to 5 , which is acquired before the application step, and a degree of change in the degree of facial vivacity after the application step;
a determining step of determining that a skin condition improving agent having a higher rate of increase in the degree of facial vitality is a superior skin condition improving agent based on the results of the measuring step;
A method for screening a skin condition improving agent , comprising: a spatial frequency calculation means for individually calculating the intensity of a spatial frequency obtained from color channel data having a specific hue among the RGB color channel data of the color face image data of the subject , and combining and using the spatial frequencies obtained from two or more types of color channel data ;
an evaluation means for evaluating the degree of lively face of the subject by applying the calculated spatial frequency intensities to an equation or model that indicates a correlation between the intensities of two or more spatial frequencies obtained from color channel data that holds a specific hue and a lively face, the equation or model being prepared in advance and indicating a correlation between the intensities of two or more spatial frequencies and a lively face;
An estimation device comprising: Computer,
a spatial frequency calculation means for individually calculating the intensity of a spatial frequency obtained from color channel data having a specific hue among the RGB color channel data of the color face image data of the subject , and combining and using the spatial frequencies obtained from two or more types of color channel data ;
an evaluation means for evaluating the lively face of the subject by applying the calculated spatial frequency intensities to an equation or model that indicates a correlation between the intensities of two or more spatial frequencies obtained from color channel data that holds a specific hue and is obtained from face image data, the intensities being prepared in advance, and the lively face;
A presumed program to function as. a storage unit that stores, as input information, a correlation between the intensities of two or more spatial frequencies obtained from color channel data that holds a specific hue among RGB color channel data of the color face image data and the degree of liveliness of the face;
A learning unit that generates a trained model in which variables for estimating a degree of a lively face of a subject from facial image data of the subject are machine-learned using the correlation as training data;
A learning device comprising: