Methods for estimating skin physiological indicators

Abstract

To provide a method for estimating a skin physiological index.SOLUTION: According to the method of estimating a skin physiological index, one cell region of a horny layer image is specified and a skin physiological index is estimated on the basis of the correlation between at least one horny layer parameter value thus calculated and the result of analysis by a facial skin image analysis counseling system.SELECTED DRAWING: None

Description

Technical Field

[0001] The present invention relates to a method for estimating skin physiological indices.

Background Art

[0002] Devices that can quantify skin physiological indices such as texture, fine wrinkles, spots, wrinkles, pores, etc. from magnified images of the skin surface or images of the face are known (Non-Patent Documents 1 and 2). Although these devices can quantify and evaluate skin physiological indices, the devices are expensive, and experts are required for image measurement and device operation. Therefore, the facilities where measurements can be made are limited to specialized salons, medical institutions, etc., and it is difficult to know skin physiological indices cheaply, simply, and quickly.

Prior Art Documents

Non-Patent Documents

[0003]

Non-Patent Document 1

Non-Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] An object of the present invention is to provide a method for estimating skin physiological indices.

Means for Solving the Problems

[0005] As a result of intensive studies to solve the above problems, the present invention has been made by finding that skin physiological indices can be estimated using the stratum corneum structure obtained from an observation image of the stratum corneum as an index.

[0006] Specifically, the means for solving the problems of the present invention are as follows: 1. Identify single-cell regions in stratum corneum images, A method for estimating skin physiological indicators, characterized by estimating skin physiological indicators based on one or more correlations selected from (1) to (8) below between the stratum corneum parameter value of 1 or more calculated therefrom and the analysis results from a facial skin image analysis counseling system. (1) Correlation between the standard deviation of single-cell brightness as the explanatory variable and the Visia Texture Count as the dependent variable. (2) Correlation between one or more stratum corneum parameters selected from the group consisting of single cell area, area of ​​the rectangle circumscribing a single cell, single cell brightness standard deviation, single cell brightness maximum value, and standard deviation of circularity in the image as explanatory variables, and Visia UV Spot Count as the dependent variable. (3) Correlation between the ratio of the long and short sides of a rectangle as the explanatory variable and the Visia Wrinkle Count as the dependent variable. (4) Correlation between one or more stratum corneum parameters selected from the group consisting of roundness, single cell area, single cell perimeter, area of ​​the rectangle circumscribing a single cell, standard deviation of single cell area in the image, and standard deviation of single cell perimeter in the image, with Brown Spot Count_Visia as the dependent variable. (5) Correlation between one or more stratum corneum parameters selected from the group consisting of roundness, standard deviation of roundness in the image, and standard deviation of single-cell brightness as explanatory variables, and Pore Count_Visia as the dependent variable. (6) Correlation between one or more stratum corneum parameters selected from the group consisting of single cell area, single cell perimeter, area of ​​a rectangle circumscribing a single cell, an approximate quadrilateral-hexagon, and the standard deviation of the single cell area in the image, and Visia UV Spots as the dependent variable. (7) Correlation between the ratio of the long and short sides of a rectangle as the explanatory variable and Visia Wrinkles as the dependent variable. (8) Correlation between one or more stratum corneum parameters selected from the group consisting of single cell area, single cell perimeter, ratio of long and short sides of a rectangle, average single cell brightness, approximate 4-6 polygon, standard deviation of single cell brightness, standard deviation of single cell area in the image, standard deviation of single cell perimeter in the image, standard deviation of minimum single cell brightness in the image, and standard deviation of average single cell brightness in the image, with Spot Count_Visia as the dependent variable. 2. The method for estimating skin physiological indicators according to 1, characterized in that the single-cell region is identified using a machine learning model that has been trained in advance with multiple training stratum corneum images and single-cell regions obtained by visual evaluation in each training stratum corneum image. [Effects of the Invention]

[0007] The present invention's method for estimating skin physiological indicators uses numerical values ​​obtained from the stratum corneum structure as indicators, and allows for the estimation of skin physiological indicators very easily and accurately. [Modes for carrying out the invention]

[0008] This invention identifies a single-cell region in a stratum corneum image, This invention relates to a method for estimating skin physiological indicators based on the correlation between a stratum corneum parameter value of 1 or greater, calculated from the present invention, and the analysis results obtained by a facial skin image analysis counseling system.

[0009] The following skin physiological indicators can be estimated using the skin physiological indicator estimation method of the present invention. Visia Texture Count: An indicator of skin smoothness. The area of ​​changes in the skin surface (raised or depressed areas). Spot Count_Visia: Brown or red spots. Spots that have an area within a certain range that has a different contrast to the background skin tone. Pore ​​Count_Visia: Number of pores. Identifies pores that are darker than the surrounding skin tone, circular in shape, and relatively small. Brown Spot Count_Visia: An indicator of surface or visible pigmentation of the skin. Brown spots that cover a certain area and have a contrast with the background skin tone. Visia UV Spots: An indicator of epidermal melanin levels. These are spots that absorb ultraviolet rays just below the skin's surface. Visia UV Spot Count: Pigmentation within the skin that may appear on the skin's surface in the future. Visia Wrinkles: The area of ​​wrinkles. Determined by the contrast with the background skin tone and the linear shape. Visia Wrinkle Count: The number of wrinkles. Determined by the background skin tone, contrast, and linear shape.

[0010] The method for collecting the stratum corneum to obtain images can be either biopsy or tape stripping, but tape stripping is preferred because it is less burdensome for the subject. Tape stripping is a method in which an adhesive tape is applied to the skin and then peeled off to collect the surface layer of the skin.

[0011] Stratum corneum cells collected by tape stripping are imaged using transmitted light observation (including differential interference contrast, phase contrast, and dark-field observation) or reflected light observation to observe the morphology of the cells. Stratum corneum cells are preferably observed in an unstained state, but they can be stained if necessary. Cell observation can be performed using a microscope capable of cell observation, such as the VHX-500 digital microscope from Keyence Corporation or the Dino-Lite digital microscope from AnMo Electronics Corporation. Observation conditions are not limited as long as they allow for detailed observation of cells, but examples include conditions with a resolution of 1.0 μm / pixel or higher and approximately 200,000 pixels or more.

[0012] The captured stratum corneum images are processed using an image processing system to identify single-cell regions within the stratum corneum images. During this process, it is also possible to identify cell region area, stratified exfoliation region, and stratified exfoliation rate. The single-cell region refers to the region of each stratum corneum cell within the stratum corneum image. The single-cell region area is the sum of the areas of the stratum corneum cell regions in the stratum corneum image. The multi-layer peeling area is the area of the region where two or more layers of the stratum corneum are peeled off overlapping each other. The multi-layer peeling rate is the ratio of the "multi-layer peeling area" to the "single-cell region area" (multi-layer peeling area / single-cell region area).

[0013] And based on this identified single-cell region, etc., according to the type of skin physiological index to be estimated, etc., one or more selected from the group consisting of the single-cell region area, the single-cell area, the standard deviation of the single-cell area in the image, the single-cell perimeter length, the standard deviation of the single-cell perimeter length in the image, the area of the rectangle circumscribing a single cell, the ratio of the long side to the short side of the rectangle, the circularity, the standard deviation of the circularity in the image, the approximated 4 - 6 polygon, the average value of the single-cell luminance, the standard deviation of the average value of the single-cell luminance, the standard deviation of the single-cell luminance, and the maximum value of the single-cell luminance are quantified. The stratum corneum image may use a color image, may use an image converted to grayscale by image processing, or both can be used.

[0014] The single-cell area is the value obtained by averaging the areas of each stratum corneum cell in the stratum corneum image over all cells. The single-cell perimeter length is the value obtained by averaging the perimeter lengths of each stratum corneum cell in the stratum corneum image over all cells. The area of the rectangle circumscribing a single cell is the value obtained by averaging the areas of the rectangles with the minimum area covering each stratum corneum cell in the stratum corneum image over all cells. The ratio of the long side to the short side of the rectangle is the value obtained by averaging the ratio of the long side to the short side (long side / short side) of the rectangle with the minimum area covering each stratum corneum cell in the stratum corneum image over all cells. The circularity is the value obtained by averaging over all cells in one image of the degree of circularity (calculated by 4π×(area) / (perimeter squared)) of the region of each stratum corneum cell. The approximate 4-6 polygon is determined by arranging regular 4-6 polygons within the stratum corneum cell region, centering on the centroid of each individual stratum corneum cell shape. The angles and sizes are adjusted so that the area of ​​the region extending beyond the regular 4-6 polygon is minimized. The area of ​​the region extending beyond the polygon is then divided by the area of ​​the stratum corneum cell. The n that minimizes this value among regular n-gons (n ​​is 4 to 6) is defined as the approximate n-gon, and the value of n for each individual stratum corneum cell in the stratum corneum image is averaged across all cells.

[0015] The average single-cell brightness is the average brightness value within the cell region of each individual stratum corneum cell in a stratum corneum image, averaged across all cells. The single-cell brightness standard deviation is the value obtained by calculating the standard deviation of brightness within the cell region of individual stratum corneum cells in a stratum corneum image. The maximum single-cell brightness is the average of the maximum brightness values ​​within the cell region of individual stratum corneum cells in a stratum corneum image, across all cells. The following values, selected from the group consisting of cell region area, single cell area, standard deviation of single cell area in the image, single cell perimeter, standard deviation of single cell perimeter in the image, area of ​​rectangle circumscribing a single cell, ratio of the long and short sides of the rectangle, roundness, standard deviation of roundness in the image, approximate quadrilateral-hexagon, average single cell brightness, standard deviation of average single cell brightness, single cell brightness standard deviation, and maximum single cell brightness, are also called stratum corneum parameter values.

[0016] The digitization using an image processing system can be performed using a known image processing system, for example, the image processing software included with the digital microscope mentioned above, or commercially available image processing software. Alternatively, an image processing system can be used that has been pre-trained with multiple stratum corneum images and a machine learning model that has been trained on single-cell regions (regions of individual cells) in each training stratum corneum image based on visual evaluation. This machine learning model can also be trained on cell regions (regions where cells exist) based on visual evaluation, multilayered exfoliation regions (regions where two or more layers of stratum corneum have been exfoliated) based on visual evaluation, etc. By inputting a stratum corneum image to be quantified into this machine learning model, outputting a single-cell region of this stratum corneum image, and calculating stratum corneum parameter values ​​(cell region area, single-cell area, standard deviation of single-cell area in the image, single-cell perimeter, standard deviation of single-cell perimeter in the image, area of ​​rectangle circumscribing a single cell, ratio of long and short sides of rectangle, roundness, standard deviation of roundness in the image, approximate 4-6 polygon, average single-cell brightness, standard deviation of average single-cell brightness, standard deviation of single-cell brightness, maximum single-cell brightness) based on the single-cell region output by the machine learning model, it is possible to quickly calculate stratum corneum parameter values ​​equivalent to those based on visual evaluation. Furthermore, it is also possible for a person to visually correct the single-cell region output by the image processing system equipped with the machine learning model to obtain stratum corneum parameter values.

[0017] The method for estimating skin physiological indicators using these quantified stratum corneum parameter values ​​as indicators is not particularly limited, and known methods can be used. For example, the stratum corneum parameter values ​​may be subjected to simple linear regression analysis or multiple linear regression analysis. The multiple linear regression analysis may be linear multiple linear regression analysis or nonlinear multiple linear regression analysis. In addition, other parameters other than cell region area, single cell area, standard deviation of single cell area in the image, single cell perimeter, standard deviation of single cell perimeter in the image, area of ​​the rectangle circumscribing a single cell, ratio of the long and short sides of the rectangle, roundness, standard deviation of roundness in the image, approximating quadrilateral-hexagon, mean single cell brightness, standard deviation of mean single cell brightness, standard deviation of single cell brightness, and maximum single cell brightness may be included as explanatory variables in the multiple linear regression analysis. [Examples]

[0018] <Sample for machine learning> Stratum corneum cells were collected from the faces of 412 women (aged 18-87, average age 46.8) using the tape stripping method. The collected stratum corneum cells were photographed unstained using a digital microscope (Keyence Corporation, VHX-5000) with transmitted light at 8-bit (RGB color), a resolution of 0.41 μm / pixel, and a pixel count of 1600 × 1200 pixels. Two to three fields of view were captured for each sample and used for machine learning.

[0019] Using the image annotation software Labelme, cell regions, stratified exfoliation regions, and single-cell regions were visually identified to obtain stratum corneum images for training. We used machine learning on each training stratum corneum image, along with cell regions, stratified exfoliation regions, and single-cell regions evaluated visually, to obtain a machine learning model capable of outputting cell regions, stratified exfoliation regions, and single-cell regions. <Sample> Stratum corneum cells were collected from the faces of 227 women (aged 28-87, average age 47.2 years) using the tape stripping method. 3-5 samples were collected from each woman, resulting in a total of 966 samples. The sample preparation method is the same as that used for machine learning samples. Sample images were analyzed using an image processing device equipped with a machine learning model, and the following parameters were quantified for each stratum corneum image: cell region area, single cell area, standard deviation of single cell area in the image, single cell perimeter, standard deviation of single cell perimeter in the image, area of ​​rectangle circumscribing a single cell, ratio of the long and short sides of the rectangle, roundness, standard deviation of single cell roundness in the image, approximate 4-6 polygon, average single cell brightness, standard deviation of average single cell brightness in the image, standard deviation of single cell brightness, and maximum single cell brightness. These parameters were then used to determine the stratum corneum parameters. Furthermore, the median single-cell brightness, minimum single-cell brightness, standard deviation of the standard deviation of single-cell brightness within the image, standard deviation of the median single-cell brightness within the image, standard deviation of the maximum single-cell brightness within the image, standard deviation of the minimum single-cell brightness within the image, and standard deviation of the approximate quadrilateral within the image were also quantified to determine the stratum corneum parameter values. The median luminance of a single cell is the average of the median luminance values ​​within the cell region of individual stratum corneum cells in a stratum corneum image, across all cells. The single-cell minimum brightness is the average of the minimum brightness values ​​within the cell region of individual stratum corneum cells in a stratum corneum image across all cells.

[0020] Furthermore, for 966 samples, single-cell regions, cell regions, and stratified delamination regions, as determined by visual evaluation, were compared with single-cell regions, cell regions, and stratified delamination regions analyzed using an image processing device equipped with a machine learning model. The results showed that the ratio of the single-cell area area obtained by visual inspection to the single-cell area area analyzed by an image processing device equipped with a machine learning model was 117%, the ratio of the single-cell area area obtained by visual inspection to the single-cell area area analyzed by an image processing device equipped with a machine learning model was 95%, and the ratio of the multilayered delamination area area obtained by visual inspection to the multilayered delamination area area analyzed by an image processing device equipped with a machine learning model was 86%. Thus, the single-cell areas, cell areas, and multilayered delamination areas obtained by visual inspection were equivalent to those obtained by an image processing device equipped with a machine learning model.

[0021] <Regression analysis of stratum corneum structure analyzed using an image processing device equipped with a machine learning model of skin physiological indicators> Skin physiological index measuring device name: VISIA evolution (manufactured by Canfield Scientific (New York, USA)) Participants: 335 people (ages 20-92, average age 44.3) Measurement site, procedure After washing their faces, the subjects rested for 10 minutes, then their entire faces were photographed using VISIA evolution, and the stratum corneum of the cheeks was collected using the tape stripping method. The collected stratum corneum was imaged using a digital microscope Dino-Lite (AnMo Electronics Corporation) under the conditions of 8-bit RGB, 0.41 μm / pixel, and 1600 × 1200 pixels. Cell shape recognition was performed using AI and binarization, and the values ​​of each stratum corneum parameter were calculated. Regression analysis was performed with each stratum corneum parameter value as the explanatory variable and the output values ​​from each VISIA, which are skin physiological indicators, as the dependent variable. JMP(registered trademark) 16.2.0 (SAS Institute Inc., NC, USA) was used for the analysis. The output values ​​from each VISIA, which are skin physiological indicators, are Visia Texture Count, Spot Count_Visia, Pore Count_Visia, Brown Spot Count_Visia, Visia UV Spots, Visia UV Spot Count, Visia Wrinkles, and Visia Wrinkle Count.

[0022] "Simple linear regression analysis" Simple regression analysis was performed on each stratum corneum parameter value and skin physiological index. For those where a correlation was found, the correlation coefficient r, p-value, and the equation of the simple regression line (Y=aX+b, where X is each stratum corneum parameter value and Y is the skin physiological index) are shown in Tables 1 and 2. [Table 1]

[0023] [Table 2]

[0024] As shown in Tables 1 and 2, each stratum corneum parameter value described in (1) to (8) of the present invention had a correlation (|r|≧0.2) with the skin physiological index described in each of (1) to (8).

[0025] Multiple regression analysis Multiple regression analysis was performed using 20 stratum corneum parameter values, including stratum corneum parameters for which no correlation was found, as explanatory variables for each skin physiological indicator. The equation of the multiple regression line is (Y=a1X1+a2X2+···+a 14 X 14 Tables 3 and 4 show the values ​​of each stratum corneum parameter (+b, Xn are the values, and Y is the amount of marker protein).

[0026] [Table 3]

[0027] [Table 4]

[0028] As shown in Tables 3 and 4, a higher correlation with skin physiological indicators was observed when the stratum corneum parameter values ​​described in (1) to (8) of the present invention were combined with other parameter values.

Claims

1. Identify single-cell regions in stratum corneum images, A method for estimating skin physiological indicators, characterized by estimating skin physiological indicators based on one or more correlation relationships selected from (1) to (8) below between the stratum corneum parameter value of 1 or more calculated therefrom and the analysis results from a facial skin image analysis counseling system. (1) Correlation between the standard deviation of single-cell brightness as the explanatory variable and the Vishia Texture Count as the dependent variable. (2) Area of ​​one cell, area of ​​the rectangle circumscribing one cell, standard deviation of brightness of one cell, maximum brightness of one cell, Correlation between one or more stratum corneum parameters selected from a group consisting of the standard deviation of roundness in the image as explanatory variables and Visia UV Spot Count as the dependent variable. (3) Correlation between the ratio of the long and short sides of a rectangle as the explanatory variable and the Vishia Wrinkle Count as the dependent variable. (4) Correlation between one or more stratum corneum parameters selected from the group consisting of roundness, single cell area, single cell perimeter, area of ​​the rectangle circumscribing a single cell, standard deviation of single cell area in the image, and standard deviation of single cell perimeter in the image, with Brown Spot Count_Visia as the dependent variable. (5) Correlation between one or more stratum corneum parameters selected from the group consisting of roundness, standard deviation of roundness in the image, and standard deviation of single-cell brightness as explanatory variables, and PoreCount_Visia as the dependent variable. (6) Correlation between one or more stratum corneum parameters selected from the group consisting of single cell area, single cell perimeter, area of ​​a rectangle circumscribing a single cell, an approximate quadrilateral-hexagon, and the standard deviation of the single cell area in the image, with Vishia UV Spots as the dependent variable. (7) Correlation between the ratio of the long and short sides of a rectangle as the explanatory variable and Vishia Wrinkles as the dependent variable. (8) Correlation between one or more stratum corneum parameters selected from the group consisting of single cell area, single cell perimeter, ratio of long and short sides of a rectangle, average single cell brightness, approximate 4-6 polygon, standard deviation of single cell brightness, standard deviation of single cell area in the image, standard deviation of single cell perimeter in the image, standard deviation of minimum single cell brightness in the image, and standard deviation of average single cell brightness in the image, with Spot Count_Visia as the dependent variable.

2. The method for estimating skin physiological indicators according to claim 1, characterized in that the single-cell region is identified using a machine learning model that has been trained in advance on multiple training stratum corneum images and single-cell regions obtained by visual evaluation in each training stratum corneum image.

Images