JP2017158921A - Skin condition evaluation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique that makes it possible to evaluate skin conditions based on a variety of indexes.SOLUTION: A skin condition evaluation method comprises an acquisition step (S13) for acquiring a blur image of a predetermined blur strength in which a skin surface is captured, and a generation step (S15) for generating distribution information in a specific brightness range in the blur image.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a skin condition evaluation technique.

One of the indices for evaluating the skin condition is texture (texture), which includes various impression elements such as gloss, transparency, beautiful skin, darkness, unevenness, and aging (age impression). Can be included.
Patent Literature 1 below discloses a method for evaluating a subject's beauty on the skin as a texture using the subject's face image. In this method, an index value related to the area or shape of a predetermined gradation area is calculated from the luminance gradation image of the cheek area of the subject's face image, and the cosmetic elasticity of the subject is evaluated based on the index value. Is done.

In addition, the skin condition can be evaluated based on the physical characteristics of the skin surface such as unevenness and color unevenness.
The following Patent Document 2 discloses a method for evaluating cheek sagging as a feature of the skin surface. Specifically, a method for evaluating the effect of correcting cheek sagging by the foundation is disclosed. In this method, the lightness distribution of the virtual face image averaged by the morphing technique is represented by lightness contour lines of a plurality of gradations, and the angle and area obtained by using the contour lines are calculated.

JP 2008-93048 A JP 2005-237409 A

  According to each method described above, it is possible to make an evaluation based on a limited index such as “harness” or “cheek sagging”. However, no method has yet been established that can achieve skin condition evaluation across various indices such as texture and shape characteristics.

  This invention is made | formed in view of said subject, and provides the technique which enables the skin state evaluation by a various parameter | index. In this specification, “evaluation of skin condition” means to evaluate the apparent condition of the skin surface for non-medical purposes, and includes evaluations that are possible even by an evaluator other than an expert.

  The skin condition evaluation method according to the first aspect of the present invention includes an acquisition step of acquiring a blur image having a predetermined blur intensity obtained by photographing a skin surface, and a generation step of generating distribution information of a specific brightness range in the blur image. And including.

  The cosmetic according to the second aspect of the present invention, a cosmetic treatment (excluding medical practice) or a method for evaluating the effect of a cosmetic or health food or drink, the cosmetic, the cosmetic treatment (excluding medical practice) Alternatively, the skin before and after application of the food or drink for beauty or health is evaluated by the skin condition evaluation method.

  According to a third aspect of the present invention, the skin condition evaluation apparatus obtains a blur image having a predetermined blur intensity obtained by photographing a skin surface, and a generation unit that generates distribution information of a specific brightness range in the blur image. And comprising.

  According to the technique provided by the present invention, it is possible to evaluate the skin condition using various indices.

It is a flowchart which shows the skin condition evaluation method (this method) of this embodiment. It is a figure which shows notionally the example of the blurring process by several different blur intensity | strength. It is a figure which shows notionally the example of the reduction gradation process applied with respect to a blurred image. It is a figure which shows the example of the image which shows the outline image and target pixel of a local area | region produced | generated from a gradation-reduced image. It is a graph of an example of change tendency information. It is a figure which shows notionally the hardware structural example of the skin image analyzer (evaluation apparatus) in this embodiment. It is a figure which shows notionally the process structural example of the skin state evaluation apparatus 10 in this embodiment. It is a figure which shows the skin image of three people. It is a graph which shows the change tendency of the number of local areas accompanying the difference in blur intensity. It is a graph which shows the change tendency of the local area area accompanying the difference in blur intensity. It is a graph which shows the change tendency of the number of standardized local regions accompanying the difference in blur intensity. It is a graph which shows the change tendency of the standardized local region area accompanying the difference in blur intensity. It is a figure which shows the example of calculation of the variation | change_quantity of the normalized local region area. It is a figure which shows the example of the moving image of the gradation image of 8 gradations.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
First, an outline of the present embodiment will be described.
The skin condition evaluation method of the present embodiment (hereinafter sometimes referred to as the present method) includes an acquisition step of acquiring a blurred image of the skin surface, and a generation step of generating distribution information of a specific brightness range in the blurred image. Including.

  The blurred image acquired in the acquisition step is an image having a predetermined blur intensity obtained by photographing at least a part of the body of the subject including the evaluation target site on the skin surface of the subject. There are no particular limitations on the evaluation target part that appears in the blurred image. For example, the evaluation target part is determined according to an evaluation index that is a target such as a sense of transparency, gloss, aging, and color unevenness. Further, the blurred image may be acquired by cutting out an image region of the evaluation target part from an image showing a wide range surrounding the evaluation target part in the human body.

The blurred image may be acquired by performing a blurring process with a predetermined blur intensity on the skin image captured by the camera. For example, a process for averaging local features (high frequency components) while maintaining global spatial information (low frequency components) included in an unblurred skin image is selected as the blurring process. Specific blurring processing methods include a smoothing filter such as a moving average filter, a noise removal filter using non-smoothing processing such as a median filter, and the like. In the present embodiment, the blurring processing method is not limited.
The “predetermined blur strength” is a predetermined blur strength and is determined according to an evaluation index or the like. For example, when the moving average filter is used in the blurring process, the number of times the moving average filter is applied corresponds to the blurring intensity. In this case, the “blurred image with a predetermined blur intensity” is acquired by repeatedly applying a moving average filter to the original skin image a predetermined number of times. Also, the greater the filter size of the moving average filter, the greater the degree of smoothing and the greater the blur intensity. Therefore, the predetermined blur intensity can be realized by setting the filter size of the moving average filter to a predetermined size.
As another example, the blurred image may be acquired by a unique function of the imaging device (camera). For example, the blurred image may be a blurred image obtained by attaching a special lens capable of realizing a predetermined blur intensity to the camera, or a blurred image or a blurred process performed by image processing software installed in the camera. It may be. In this case, the blurred image may be acquired directly from the camera, or may be acquired indirectly from the camera via a portable recording medium or a personal computer (PC).

The blurred image may be a color image or a grayscale image of a specific color channel. Examples of the grayscale image include a pale grayscale image (grayscale image), R (red), G (green), B (blue), or a grayscale image of a specific wavelength. The grayscale image may be obtained by converting a color image by a known method such as a simple average method.
The file format of the blurred image is arbitrary, such as JPEG, GIF, TIFF, and BMP, but is preferably an uncompressed format such as TIFF and BMP. This is because loss of image information due to compression can be suppressed.

  The “distribution information” generated in the generation process is information indicating the distribution of pixels (hereinafter also referred to as target pixels) having brightness within a specific brightness range in the blurred image. The “distribution information” includes the number of target pixels (area), the number or area of local regions (islands) formed by a series of one or more target pixels, the direction indicated by a series of two or more target pixels, the target pixels One or more index values such as the extent or position of the distribution range are exemplified. As another example of “distribution information”, an image (contour image or the like) that represents the local region in the blurred image by contour lines or color coding, or an image of the distribution of the target pixel in the blurred image (such as a diagram, table, or graph) It may be.

“Specific brightness range” means a specific range related to brightness. The specific range can be determined according to the evaluation target and the brightness distribution of the blurred image. For example, when a bright component such as gloss is to be evaluated, a range with high brightness is set, and when a dark component such as a shadow is to be evaluated, a range with low brightness is set.
“Brightness” means the degree of brightness. In this specification, “lightness” is used in a broad sense including the meaning of “luminance” in addition to the general meaning of “lightness”. That is, the “lightness” in this specification may be indicated by the value of the brightness (Value or Brightness) component of the HSV or HSB model, or may be indicated by the value of the brightness (Lightness) component of the HSL model. For example, “lightness” is at least one of a maximum value, R value, G value, and B value among pixel values (R value, G value, and B value), or at least R value, G value, and B value. It is indicated by the luminance value calculated from the two. The luminance value may be a pixel value of a gray scale image (a value indicating the degree of shading).

  Here, the “local region” is a partial region formed by connecting one or more target pixels in the blurred image. A plurality of methods can be considered as a method for setting the “local region”. For example, in one method, the “local region” is set so that the “local region” is formed of one or more target pixels and adjacent target pixels belong to the same “local region”. In another method, a predetermined number or more of target pixels may be used as a condition for forming the “predetermined region”. When the predetermined number is set to 3 in this method, one target pixel having no adjacent target pixel is not regarded as a “local region”. In yet another method, a sequence of target pixels in which one or more target pixels that are not adjacent to a pixel other than the target pixel exist may be used as the “local region” formation condition.

As will be described later, the present inventor performed sensory evaluation on the appearance of the skin surface for a plurality of skin images. In the sensory evaluation, evaluation was performed with a plurality of indices such as glossiness and aging. As a result, it was found that the difference in the distribution of the specific brightness range in the skin image blurred with a predetermined intensity shows a correlation with the difference in the apparent state of the skin surface at each index.
In the present embodiment, based on such knowledge, distribution information of a specific brightness range in the blurred image on the skin surface is generated, and the skin state is evaluated based on the generated distribution information.
The blurring of the skin image at a predetermined intensity selectively makes it difficult to selectively view certain non-uniform skin elements that obstruct the uniform pale glossy impression that beautiful skin exhibits. This is because the blur strength corresponds to the visual distance of the skin, and the appearance of the skin state changes according to the visual distance. Then, by blurring the skin image, it is possible to bring it closer to an image that is uniform and can be equated with beautiful skin. Here, skin characteristics such as pores, pore dirt, wrinkles, color unevenness (stains, etc.), and shine are applicable to the non-uniform skin elements. In other words, if the blur is weak, the non-uniform skin element of fine shape and delicate contrast (high frequency component) is erased, but the non-uniform skin element of the rough shape and contrast (low frequency component) remains, and if the blur is strong, The rough non-uniform skin element is also erased. Therefore, it can be said that the skin image blurred at a predetermined intensity represents the remaining skin features from which a specific non-uniform skin element has been erased.
In the present embodiment, such a skin feature is visualized by a distribution of a specific brightness range. Thereby, for example, according to the distribution information generated from the blurred image of a certain test skin, it is possible to objectively evaluate the appearance state of the skin with a plurality of indices relating to the texture and shape characteristics.
Thus, according to this embodiment, the apparent state of the skin surface can be objectively evaluated with various indices.

[Skin condition evaluation method]
Hereinafter, this method will be described in more detail with reference to FIG.
FIG. 1 is a flowchart showing a skin condition evaluation method (this method) of the present embodiment. As shown in FIG. 1, the method includes steps (S11) to (S17).

Step (S11) is a step of acquiring a skin image of the subject.
The skin image acquired in the step (S11) is an image in which at least a part of the human body of the subject including the evaluation target site on the skin surface of the subject is captured by the imaging device (camera). The skin image acquired in the step (S11) may be described as an original skin image. The imaging device may be a general RGB camera, a monochrome camera, or a spectrum camera, and the performance and specifications of the imaging device are not limited. There are no restrictions on the shooting environment conditions such as the number of lights, lighting angle, illuminance, and shooting angle, and the appearance characteristics of the skin that affect evaluation indices such as pores, wrinkles, color unevenness, and gloss can be visually confirmed. It only has to be an environment.
The resolution of the acquired skin image may be anything as long as the above-described skin appearance characteristics are visible. The file format of the skin image is arbitrary, such as JPEG, GIF, TIFF, BMP, etc., but is preferably an uncompressed format such as TIFF or BMP. This is because loss of image information due to compression can be suppressed.
The skin image may be a color image or a grayscale image of a specific color channel. Examples of the gray image include a light gray image (grayscale image), R, G, B, or a gray image having a specific wavelength. The grayscale image may be obtained by converting a color image by a known method such as a simple average method.

Step (S12) is a step of extracting an evaluation target part from the original skin image acquired in step (S11). The evaluation target part is arbitrarily determined according to the target evaluation index or the like. Examples of the evaluation target part include at least a part of a face such as a cheek, forehead, mouth, and corner of the eye, at least a part of decollet that is a region from the neck to the chest, and at least a part of an arm and a leg.
In the step (S12), the skin image of the evaluation target part can be extracted by trimming the original skin image. At this time, you may perform the further process of expansion etc. with respect to the skin image of an evaluation object site | part at this time. In addition, an image obtained by painting a part other than the evaluation target part of the original skin image with a specific color such as black may be used as the skin image of the evaluation target part. Hereinafter, the skin image of the evaluation target part extracted in the step (S12) may be referred to as a target skin image.
The step (S12) may be omitted depending on the reflection aspect of the evaluation target part in the original skin image acquired in the step (S11). For example, when only the evaluation target part is shown in the original skin image, or when the evaluation target part occupies most, the step (S12) may not be performed.

  Step (S13) is a step of applying blurring processing with a plurality of different blurring intensities to the target skin image extracted in step (S12). Although the specific method of the blurring process is not limited, as described above, the blurring process using a blurring filter such as a smoothing filter and a noise removal filter is exemplified. Specific examples of the smoothing filter include a Gaussian filter, a moving average filter, a weighted average filter, and a low-pass filter. Depending on the characteristics of each filter, it is possible to use different filters for blurring. According to the Gaussian filter, since the attenuation pattern of the internal reflection light peculiar to beautiful skin is close to a Gaussian curve, the original skin image can be processed in a direction approaching the apparent state of the beautiful skin. Further, according to the moving average filter, it is possible to speed up the blurring process compared to other filters.

  “Blurring processing with a plurality of different blurring intensities” is realized, for example, by applying a smoothing filter to the target skin image a plurality of times. In order to increase the processing speed, it is particularly desirable to apply a moving average filter having a predetermined size a predetermined number of times. The blurring process can also be realized by applying a plurality of filters having different sizes to the target skin image.

FIG. 2 is a diagram conceptually illustrating an example of blurring processing using a plurality of different blurring intensities. FIG. 2 shows the target skin image and six blurred images having different blur strengths, and the blur strength of the blur image increases in the direction of the arrow. In the example of FIG. 2, a blurred image obtained by repeatedly applying a moving average filter of 5 pixels vertically and 5 pixels horizontally to a target skin image of 1032 pixels vertically and 770 pixels horizontally is shown.
As illustrated in FIG. 2, a plurality of blurred images having different blur strengths can be obtained by the step (S13).

  In the present embodiment, the number of blur images to be acquired, in other words, the number of blur strengths of the blur processing to be applied is two or more. By using a blurred image having two or more different blur strengths, it is possible to capture the tendency of the blur image to change as the blur strength changes. However, as described above, even if one blur image having a predetermined blur strength is used, a sufficiently effective evaluation can be performed. The verification of these effects will be described later.

Step (S14) is a step of specifying each pixel in a specific brightness range (hereinafter referred to as a target pixel) in each blurred image obtained in step (S13).
The definitions of “specific brightness range” and “lightness” are as described above.
The brightness of each pixel is indicated by a brightness value or a luminance value obtained from the pixel value of each pixel forming the blurred image. For example, the pixel value (RGB value) of each pixel is converted into an HSV model, HSB model, or HSL model by a known method, and the value of the lightness (Value or Brightness) component or luminance (Lightness) component of the obtained model is Used as the brightness value of the pixel. The brightness value of each pixel may be the maximum value among the R value, G value, and B value of each pixel forming the blurred image, or may be selected in advance from the R value, G value, and B value. The brightness value may be obtained by a known brightness value calculation method from at least two of the R value, the G value, and the B value. In this embodiment, in order to evaluate the appearance of the skin state, it is desirable to use a G value or luminance value with high human visual angle sensitivity as the brightness value.

  In the present embodiment, the “specific brightness range” is set to a range of brightness values selected in advance from a predetermined brightness value range, excluding a predetermined range including an upper limit value of brightness that can be indicated by each pixel of the blurred image. The For example, when a pixel value (luminance value) of an 8-bit grayscale image is used as a lightness value, a predetermined lightness value range excluding a predetermined range (for example, 253 to 255) including the upper limit value (255) of the lightness value ( A range of brightness values selected in advance from 0 to 252) (for example, 220 to 252) is used as the specific brightness range. As described above, by removing the predetermined range including the upper limit value of the brightness, the noise component in the image can be removed, and highly accurate evaluation can be performed. Hereinafter, a range of excluded brightness values may be referred to as an excluded brightness range. As described above, the “specific brightness range” may be set as a range of brightness values selected according to the evaluation index from the remaining brightness value ranges excluding the excluded brightness range. Since it is preferable to pay attention to glossy pixels in the skin beautification evaluation, it is preferable to set the range of the highest brightness value as the “specific brightness range” among the remaining excluded brightness ranges.

Step (S14) is performed directly or indirectly on the blurred image obtained in step (S13). As an example of indirect implementation, the step (S14) may be performed on the converted grayscale image by converting the blurred image, which is a color image, into a grayscale image of a specific color channel.
In step (S14), a process of reducing (compressing) the number of gradations of the blurred image or its grayscale image (hereinafter, sometimes referred to as a reduced gradation process) is performed, and the obtained floor reduction is performed. You may implement with respect to a tone image. In this case, the position of the pixel having the gradation value corresponding to the “specific brightness range” is specified from the reduced (compressed) gradations. For example, when a process of reducing (compressing) 256 gradations to 8 gradations is performed, a lightness value (7) corresponding to a “specific brightness range” from lightness values (0 to 7) of 8 gradations. The position of the pixel is specified.
The gradation reduction process is applied to the pixel value of the blurred image or the brightness value indicated by each pixel of the blurred image. Further, the gradation reduction process may be applied to the remaining brightness value range excluding the excluded brightness range.

  FIG. 3 is a diagram conceptually illustrating an example of the gradation reduction process applied to the blurred image. In the example of FIG. 3, the G value is used as the brightness in the color blurred image (left image). Decreasing the gradation of the G value (brightness value) of each pixel of the blurred image as a pixel value that is obtained by compressing the gradation from 253 gradations of 0 to 252 excluding the excluded lightness range to 8 gradations A toned image (right image) is generated.

Step (S15) is a step of generating distribution information for each blurred image having a different blur intensity based on the pixel number information or position information of the target pixel specified for each blurred image in step (S14). Although the definition and examples of “distribution information” are as described above, the step (S15) can calculate the following index values as distribution information. Hereinafter, an index value calculated as distribution information may be referred to as a distribution index value.
For example, in the step (S15), using the pixel number information of the target pixel, the total number of target pixels can be calculated for each blurred image having a different blur intensity as distribution information for each blurred image having a different blur intensity.
Moreover, a process (S15) can also calculate the number or area of a local area | region as distribution information. The definition and setting method of the “local region” is as described above. In the step (S15), the local region in the blurred image is detected according to the above-described local region setting method, and the number of detected local regions may be counted. In addition, the total number of target pixels forming each detected local region may be calculated as the area of the local region. When all target pixels belong to any local region, the total number of target pixels and the area of the local region are the same.
Further, the step (S15) specifies a pixel that forms a contour line of the local region based on the position information of the target pixel, and calculates a tangent direction of the contour line, thereby indicating a value for each local region. Can also be calculated respectively. In this case, a value indicating the direction calculated for each local region or a statistical value (average, variance, etc.) of the value can be used as the distribution information. Further, the statistical value in the adjacent direction of the adjacent target pixel pair may be calculated as distribution information.

  When the index value is calculated as the distribution information in this way, the step (S15) may further normalize the calculated index value and use the normalized index value as the distribution information (distribution index value). In the step (S15), the index value calculated as described above is normalized based on the distribution index value of the target pixel extracted from the unblurred skin image of the same subject. In this case, by executing the steps (S14) and (S15) for the target skin image (unblurred skin image) extracted in the step (S12), the target pixel of the target pixel is related to the target skin image. A distribution index value is generated. For example, normalization is realized by calculating a ratio of the distribution index value obtained from the blurred image with respect to the distribution index value of the target skin image. By normalizing the distribution information in this way, the distribution information obtained from other skin images and the standard are unified, so that it is easy to compare with other skin, and therefore, the skin condition evaluation is facilitated. Can do. In addition, since the standardization is performed by the same person, the influence of shooting conditions can be eliminated.

In the step (S15), the following image can be generated as distribution information.
For example, the step (S15) can generate, as distribution information, an image that can distinguish the target pixel from the other pixels from each blurred image, its grayscale image, or its reduced gradation image. As an example, an image (see FIG. 4) that represents the target pixel and other pixels with binary pixel values can be generated.
In the step (S15), an image (see FIG. 4) representing the local region in the blurred image by a contour line or color coding can be generated as distribution information.
FIG. 4 is a diagram illustrating an example of a contour image of a local region generated from a reduced gradation image and an image showing a target pixel. In FIG. 4, the target pixel having the lightness in a specific lightness value range is the pixel having the highest pixel value in the reduced gradation image, and the local region is formed by linking one or more such target pixels. An example of a partial area is shown. In the contour image exemplified in FIG. 4, the contour line of the local region is shown in black, and in the image showing the target pixel, the target pixel is shown in black and the other pixels are shown in white. ing. In the example of FIG. 4, two local regions are detected.

  Further, the reduced gradation image shown in FIG. 3 may be generated as distribution information. When a certain gradation value in the reduced gradation image corresponds to the specific brightness range described above, the reduced gradation image is an image that can distinguish the target pixel from other pixels. In addition, when all target pixels belong to any local region, the reduced gradation image can also be referred to as an image representing the local region.

In the step (S16), information indicating a change tendency accompanying a difference in blur intensity in a plurality of distribution information respectively generated for each blur image having different blur intensity (hereinafter, sometimes referred to as change trend information) is generated. It is a process.
FIG. 5 shows a graph of an example of the change tendency information. As illustrated in FIG. 5, in the step (S 16), a change trend graph having the blur intensity and the distribution index value as coordinate axes can be generated as change trend information. The change trend graph is not limited to the example of FIG. For example, the vertical axis and the horizontal axis may be interchanged. Further, on the horizontal axis illustrated in FIG. 5, the blur intensity is arranged in ascending order, but the blur intensity may be arranged in descending order. FIG. 5 shows a graph in which the calculated distribution index values are interpolated with curves, but the distribution index values may be linearly interpolated, or only the calculated distribution index values are indicated by discrete points. May be.

As another example of the change trend information, the step (S16) can generate a change trend table that represents the relationship between the blur intensity and the distribution index value in a tabular form.
In the step (S16), the change amount or change rate of the distribution index value according to the difference in blurring intensity, the average or derivative value thereof can be calculated as the change tendency information. For example, the change amount, the change rate, and the differential value may be calculated with respect to a predetermined blur intensity.
Further, in the step (S16), when an image for each blurred image having different blur strength is generated as distribution information for each blur image having different blur strength, a moving image in which the images are arranged in ascending or descending order of the blur strength. Can also be generated as change tendency information.
The change tendency information only needs to indicate a change tendency associated with a difference in blur intensity among a plurality of distribution information, and a specific mode thereof is not limited.

  Step (S17) is a step of evaluating the skin condition of the subject. The step (S17) can be executed using either one or both of the distribution information generated for each blurred image having a different blur intensity in the step (S15) and the change tendency information generated in the step (S16). When the change tendency information is not used in the step (S17), the step (S16) may be omitted. In the present embodiment, the specific evaluation method in the step (S17) is not limited.

For example, the subject himself or a third party becomes an evaluator, and the evaluator compares the distribution information for each blurred image with different blurring intensity generated in the step (S15) with the sample distribution information, thereby determining the skin of the subject. (Test skin) can be evaluated.
In this case, distribution information for each of the blurred images having different blurring intensities relating to the sample skin is generated in advance by the above-described steps (S11) to (S15), and the generated sample distribution information is held in some form. . The sample distribution information may be stored in a computer or a portable recording medium, or may be printed out on a paper medium. Further, as the sample skin from which the sample distribution information is based, it is desirable that a representative skin is selected for each of a plurality of evaluation indices. For example, at least one skin having at least one negative impression and ideal skin such as skin with strong aging, skin with strong shine, skin with strong color unevenness, and skin with conspicuous pores is selected. In addition, skin with a high possibility that the subject desires comparison may be selected, such as average skin for each age or gender, skin of a celebrity (actress, model, singer, etc.). The average skin can be generated by combining sample skin images of a plurality of people.

The evaluator can also evaluate the state of the test skin by comparing the change tendency information generated for the test skin with the change trend information generated for the sample skin. The change tendency information regarding the sample skin may be generated in advance and held in some form.
The evaluator compares the distribution information of the subject with the sample distribution information or compares the change tendency information of the test skin with the change tendency information of the sample skin for each blurred image having different blur intensity. The evaluator evaluates the skin condition of the subject by determining whether the subject's skin is close to the sample skin, whether the subject's skin is close to the sample skin, or which sample skin is close to the subject's skin. Can do.

Here, each step of the method may be executed by a person operating a computer. For example, in the step (S11), a person may take a skin image of the subject by operating the camera, and the person may operate the computer to capture the skin image in the computer. Step (S12) and step (S13) can be performed by a person operating a commercially available image editing software. A process (S14) and a process (S15) can also be performed by manipulating commercially available software. As described above, the evaluator (person) can execute the step (S16) without using a computer.
Moreover, each process of this method may be automatically performed by a computer (for example, a skin condition evaluation apparatus described later). Step (S16) may be executed by a computer such as a skin condition evaluation apparatus described later.

[Skin condition evaluation device]
Hereinafter, the skin condition evaluation apparatus 10 which implement | achieves this method is demonstrated.
FIG. 6 is a diagram conceptually illustrating a hardware configuration example of the skin image analysis apparatus 10 in the present embodiment. A skin condition evaluation apparatus (hereinafter sometimes abbreviated as an evaluation apparatus) 10 according to the present embodiment is a so-called computer. For example, a CPU (Central Processing Unit) 11, a memory 12, and an input connected to each other via a bus. An output interface (I / F) 13 and a communication unit 14 are included. The number of hardware elements forming the evaluation device 10 is not limited, and these hardware elements can be collectively referred to as an information processing circuit.

In addition to a general CPU, the CPU 11 includes an application specific integrated circuit (ASIC), a DSP (Digital Signal Processor), a GPU (Graphics Processing Unit), and the like.
The memory 12 is a RAM (Random Access Memory), a ROM (Read Only Memory), or an auxiliary storage device (such as a hard disk).
The input / output I / F 13 can be connected to user interface devices such as the display device 15 and the input device 16. The display device 15 is a device that displays a screen corresponding to drawing data processed by the CPU 11 or the like, such as an LCD (Liquid Crystal Display) or a CRT (Cathode Ray Tube) display. The input device 16 is a device that receives an input of a user operation such as a keyboard and a mouse. The display device 15 and the input device 16 may be integrated and realized as a touch panel.
The communication unit 14 performs communication via a communication network with other computers, exchange of signals with other devices, and the like. A portable recording medium or the like can be connected to the communication unit 14. For example, an imaging device (camera) (not shown) may be connected to the communication unit 14. In this case, the camera may capture a skin image (still image) and send the still image data to the evaluation device 10.
The evaluation device 10 may include hardware elements not shown in FIG. 6, and the hardware configuration of the evaluation device 10 is not limited.

FIG. 7 is a diagram conceptually illustrating a processing configuration example of the skin condition evaluation apparatus 10 in the present embodiment.
The evaluation apparatus 10 according to the present embodiment includes an image acquisition unit 21, a blur processing unit 22, a generation unit 23, an information storage unit 24, an output processing unit 25, and the like. These processing modules are software elements, and are realized, for example, by executing a program stored in the memory 12 by the CPU 11. This program may be installed from a portable recording medium such as a CD (Compact Disc) or a memory card or another computer on the network via the input / output I / F 13 and stored in the memory 12.

The image acquisition unit 21 acquires a skin image obtained by photographing the skin surface. Specifically, the image acquisition unit 21 captures image data of a subject's skin image captured by an imaging device (camera) from the imaging device, a portable recording medium, or another computer by communication. Further, the image acquisition unit 21 may be a built-in camera of the evaluation device 10.
The image acquisition unit 21 may extract the skin image of the evaluation target part by trimming image data acquired from the outside as necessary. That is, the image acquisition unit 21 can execute the step (S11) and the step (S12) of this method. In addition, the image acquisition unit 21 may apply gray scale to the captured image data.
Hereinafter, the image data acquired by the image acquisition unit 21 may be referred to as a target skin image.

The blur processing unit 22 performs a blur process on the target skin image acquired by the image acquisition unit 21, thereby acquiring a plurality of blur images having different blur intensities. The blur processing unit 22 has a blur filter such as a smoothing filter and a noise removal filter in advance. For example, the blurring processing unit 22 acquires a plurality of blurred images having different blurring strengths by applying a smoothing filter having a predetermined size to the target skin image a plurality of times. That is, the blurring processing unit 22 can execute the step (S13) of this method.
The image acquisition unit 21 and the blur processing unit 22 correspond to an acquisition unit that acquires a blur image with a predetermined blur intensity obtained by photographing the skin surface.

The generation unit 23 corresponds to a generation unit that generates distribution information of a specific brightness range in the blurred image acquired by the blur processing unit 22. When the blur processing unit 22 generates a plurality of blur images having different blur strengths, the generation unit 23 generates distribution information for each blur image having different blur strengths. For example, the generation unit 23 uses, as the distribution information, an index value indicating at least one of an area, the number of local regions, and a direction based on pixels in a specific brightness range in the blurred image, or a local region in the blurred image. Generate an image to represent.
The “specific brightness range” is preferably a range of brightness values selected from a predetermined brightness value range excluding a predetermined range including an upper limit value of brightness that can be indicated by each pixel of the blurred image. Such “specific brightness range”, “distribution information”, “local region”, and the like are as described above.

Similarly, the generation unit 23 may generate distribution information of a specific brightness range regarding the target skin image that is not blurred and acquired by the image acquisition unit 21.
Furthermore, the generation unit 23 may also generate distribution information regarding the sample skin. The sample skin is as described above.
The generation unit 23 stores the generated distribution information in the information storage unit 24. Specifically, the generation unit 23 stores identification information for identifying the blur intensity (including identification information without blur) and distribution information in the information storage unit 24 in association with each other. Moreover, the production | generation part 23 may further associate and store the discrimination information for distinguishing sample skin and test skin with distribution information.

  The information storage unit 24 stores the distribution information generated by the generation unit 23 in a state associated with the blur intensity identification information. When storing the distribution information of the sample skin, the information storage unit 24 generates the distribution information in a state in which the above-described discrimination information is further associated. Further, when the evaluation device 10 targets a plurality of subjects, the information storage unit 24 can also store distribution information in a state of being further associated with the identification information of the subjects.

When the index value is calculated as the distribution information, the generation unit 23 standardizes the calculated distribution index value based on the distribution information (distribution index value) of the specific brightness range generated from the skin image without blurring. May be used. At this time, the generation unit 23 may extract the distribution information of the skin image that is not blurred from the information storage unit 24. The distribution index value normalized here is also expressed as distribution information. The normalization method is as described above.
Thus, the production | generation part 23 can perform the process (14) and process (S15) of this method.

Furthermore, the generation unit 23 may further generate information (change trend information) indicating a change tendency associated with a difference in blur intensity in a plurality of pieces of distribution information generated for each blur image having different blur intensity. At this time, the generation unit 23 can extract and use the distribution information for each blurred image with different blur intensity from the information storage unit 24. The change tendency information is also as described above.
Thus, the production | generation part 23 can also perform the process (S16) of this method.

The output processing unit 25 outputs one or both of the distribution information and the change tendency information generated by the generation unit 23 (hereinafter may be referred to as output information). The output form of the output information is not limited. The output processing unit 25 may display and output the output information on the display device 15, print it on a printing device (not shown), or output it as a file to a portable recording medium or another computer. Also good. Further, when the output information is a numerical value, the output processing unit 25 can also output the output information by voice.
Moreover, when the distribution information regarding the sample skin is stored in the information storage unit 24, the output processing unit 25 can output the sample distribution information for comparison with the distribution information of the test skin.

Furthermore, the output processing unit 25 may evaluate the skin condition of the subject and output the evaluation result. The output processing unit 25 evaluates the skin condition of the subject by comparing the sample distribution information with the distribution information about the test skin, or comparing the change trend information about the sample skin with the change trend information about the test skin. Distribution information and change tendency information to be compared are stored in the information storage unit 24.
For example, the output processing unit 25 calculates the difference between the distribution index value of the subject and the distribution index value of the sample skin or the difference between the change tendency information of the test skin and the change tendency information of the sample skin for each blurred image having a different blur intensity. calculate. The output processing unit 25 outputs the calculated difference value as a proximity score (evaluation result) with the sample skin. At this time, the output processing unit 25 can also output a relative table (closeness score comparison table) between each of the plurality of sample skins and each difference value.
In addition, the output processing unit 25 evaluates whether or not the calculated difference value is smaller than a predetermined threshold value, based on an evaluation result message (for example, “Your skin is beautiful skin”). Can also be output. In addition, the output processing unit 25 compares the calculated difference value between the sample skins, thereby evaluating an evaluation result message (for example, “Your skin is beautiful skin” indicating that the sampled skin is the smallest difference value. Is a little farther from aging skin and shine skin. ”) Can also be output.

The evaluation device 10 can also acquire information specifying the sample skin to be compared in response to a user operation on the input device 16. In this case, the evaluation device 10 may cause the display device 15 to output a display for designating a desired sample skin from among a plurality of sample skins whose distribution information is stored in the information storage unit 24.
Further, the evaluation device 10 may acquire evaluation information or designation information of a specific brightness range in accordance with a user operation on the input device 16. At this time, the user performs an operation of designating an evaluation index or a specific brightness range from the display output on the display device 15 according to the subject's skin imaging environment, the evaluation index, and the like.

As described above, in the present method and the evaluation apparatus 10, by performing blurring processing on the skin image in which the skin surface of the subject is reflected, a plurality of blurred images having different blur strengths are acquired, and the blur strengths are different. For each blurred image, distribution information of a specific brightness range in the acquired blurred image is generated.
According to a plurality of blurred images having different blur intensities, it is possible to reproduce on the image changes in the appearance of skin features accompanying changes in the viewing distance. In other words, according to the distribution information for each blurred image with different blurring intensity, the presence / absence, degree, and characteristics (such as high frequency component or low frequency component) of non-uniform skin elements that inhibit uniform skin characteristics of beautiful skin are grasped. be able to. Taking the area of the local area as an example of the distribution information as an example, if a high brightness range is set as a specific brightness range, if the area does not change so much due to the difference in blurring intensity, it is not good for the skin. It can be confirmed that there are few uniform skin elements. Further, when the blur intensity decreases from a certain value, when the area of the local region increases rapidly, it can be confirmed that the skin contains many non-uniform skin elements of low frequency components. .
According to the present embodiment, by using the distribution information of the specific brightness range for each blurred image with different blur strength, the presence / absence, degree, characteristic, etc. of the non-uniform skin element can be grasped and influenced by them. The apparent state of the skin can be easily evaluated with high accuracy.

For the purpose of skin care or improving the appearance of the skin, various cosmetics, beauty treatments, beauty or health foods and drinks, and the like are used. Cosmetics are applied to the skin and include, for example, lotion, milky lotion, cosmetic liquid, cream, foundation and the like. The cosmetic treatment is a treatment for the skin such as a technique for improving dullness and acne scars and a technique for improving fine lines and dents, and is a non-medical treatment. Beauty and health foods and drinks are foods and drinks for skin care, and products containing collagen, hyaluronic acid and the like are in circulation.
Therefore, by using the skin image of the subject's skin before and after the application of cosmetics, beauty treatment, or cosmetic or health food or drink, the state of the skin shown in each image is evaluated by the above-described method or the evaluation device 10. By evaluating, the improvement effect of the skin state regarding the said cosmetics, the said beauty treatment, or the said food / beverage products can be quantitatively evaluated. Specifically, this evaluation method generates the distribution information from the skin image of the subject before applying the cosmetic, the cosmetic treatment, or the food and drink, and further, the skin information of the subject after the application. Generate distribution information. By comparing the generated distribution information, it is possible to evaluate whether or not the apparent state of the skin surface is improved for the evaluation index, or the degree of improvement.

Hereinafter, the contents and results of the experiment conducted by the inventor will be described, and the effects of the above-described embodiment will be verified. The contents of the experiment will be described along this method.
In this experiment, VISIA-CR (registered trademark) manufactured by Canfield in the United States was used as an imaging device, and the faces of a plurality of subjects were photographed from obliquely forward under environmental conditions in which illumination was performed from different three directions. (S11).
Next, by trimming each photographed face image, an image (skin image) of the cheek skin surface, which is the evaluation target part, was extracted (S12). In this experiment, a skin image having a color of 1217 pixels long and 835 pixels wide was extracted.
And sensory evaluation regarding the appearance state of the skin surface was performed for the extracted skin image.
FIG. 8 shows skin images of three persons who showed typical results with respect to the indicators of glossiness, aging and beautiful skin in sensory evaluation. The skin image of the subject # 1 is a skin image evaluated as “beautiful and soft gloss”. The skin image of the subject # 2 is a skin image evaluated as “a glossiness that feels aging and feels shining”. The skin image of the subject # 3 is a skin image evaluated as “a gloss feeling that feels no aging”.

  Subsequently, the moving average filter of 5 pixels in the vertical direction and 5 pixels in the horizontal direction are repeatedly applied 50 times to the skin images of the three persons shown in FIG. Each image was generated (S13). By using this moving average filter, it is possible to acquire blur images with different blur intensities at high speed. The upper limit (50 times) of the application number of the moving average filter was set to such an extent that the skin feature affecting the appearance of the skin could not be visually recognized.

Next, a pixel having the G value of the pixel value of each blurred image as the brightness value of the pixel and having a specific brightness range of 220 to 252 is specified as the target pixel (S14). In the sensory evaluation of this experiment, since an evaluation index related to a brightness component such as gloss or luster is used, the highest brightness range (220 to 252 or less) in the brightness range excluding the excluded brightness range (253 to 255). ) Is selected.
Then, based on the identified target pixel, the number and area of the local regions are calculated for each blurred image (S15). In this experiment, a local region setting method in which all target pixels belong to one of the local regions is used, and the calculation of the step (S14) and the step (S15) is performed by the density displacement measurement software (Gray-val ).

In this experiment, the graphs shown in FIGS. 9 and 10 were respectively generated as the change tendency information.
FIG. 9 is a graph showing a change tendency of the number of local regions with a difference in blur intensity. FIG. 10 is a graph showing a change tendency of a local area area with a difference in blur intensity. In FIG. 9, the vertical axis represents the number of local regions, and the horizontal axis represents the number of filter applications corresponding to the blur intensity. In FIG. 10, the vertical axis represents the area of the local region, and the horizontal axis represents the number of filter applications corresponding to the blur intensity.
According to FIG. 9, subject # 1 evaluated as so-called beautiful skin has a smaller change compared to other subjects, and subject # 2 that feels aging and shine is less than a certain blur intensity. The number of local regions is increasing rapidly. And about subject # 3 which does not have a feeling of aging but can feel shine, although the change is larger than subject # 1, the amount of increase is not larger than subject # 2.
In FIG. 10, about subject # 1, the change is small compared with another test subject like FIG. For subject # 2, there is not much difference from the change width of subject # 1, but the local area is generally smaller than subject # 1. For subject # 3, the range of change is the largest, and the local area increases as subject # 1 reverses as the blur intensity decreases.
Thus, as illustrated in FIG. 9 and FIG. 10, it is proved that the change tendency information shows a specific change tendency according to the apparent state of the subject's skin. It is demonstrated that the appearance of the skin can be evaluated with respect to multiple indicators.

Furthermore, in this experiment, the number and area of the local regions were calculated from the skin image without blurring, and the number of local regions and the area calculated for the subject were normalized based on them. In this normalization, the ratio of the number of areas and the area of the subject to the number of areas and the area of the unblurred skin image was calculated as a percentage.
FIG. 11 is a graph showing the change tendency of the normalized number of local regions accompanying the difference in blur intensity. FIG. 12 is a graph showing a change tendency of the standardized local region area with a difference in blur intensity.
According to this standardization, values between subjects are standardized based on values calculated from unblurred skin images, so that a change tendency according to skin characteristics is more remarkably expressed.
Furthermore, as indicated by arrows in FIG. 13, the amount of change, change rate, or differential value of the local region area at the specific blur intensity corresponding to the evaluation index is used as a value indicating the change tendency of the normalized local region area. Can also be calculated. FIG. 13 is a diagram illustrating a calculation example of the change amount of the normalized local region area.

In this experiment, a grayscale image having the G value of the blurred image as a luminance value is generated from each blurred image, and a luminance value of 255 or less and 253 or more is used for the grayscale image using the luminance value as a luminance value. A gradation reduction process was performed to compress the brightness value range of 253 gradations to 8 gradations except for. As a result, an 8-tone gray image was generated for each blurred image, and a moving image in which the 8-tone gray images were arranged in time series in descending order of the blur intensity was generated for each of the three skin images.
FIG. 14 is a diagram illustrating an example of a moving image of an 8-tone gray image. FIG. 14 shows four 8-tone gray images generated for four blur intensities. In this experiment, 50 8 tone gray images were generated and their moving images were generated. By playing such a video, it is possible to visually recognize changes in the distribution of a specific brightness range due to differences in blurring intensity, and what kind of texture and morphological features the skin is close to Qualitatively evaluating the subject's skin condition can be evaluated with a plurality of evaluation indices.

  The embodiments and examples of the present invention have been described above with reference to the drawings. However, these are examples of the present invention, and various configurations other than the above can also be adopted. Further, the present invention is not limited to the above-described embodiments and examples, and includes various modifications and improvements as long as the object of the present invention is achieved. Moreover, the component of the above-mentioned evaluation apparatus 10 does not need to exist independently. Multiple components are formed as a single component, one component is formed of multiple components, one component is part of another component, one component And a part of other components are allowed to overlap. Moreover, the execution order (illustrated in FIG. 1) of the processes used in the description of the above-described skin condition evaluation method is not limited to that order. The execution order of each process can be changed within a range that does not hinder the contents, and some or all of the execution timings of the plurality of processes may overlap each other.

  For example, the skin image of the subject imaged by the imaging device (camera) may be normalized, and the normalized skin image may be blurred. In this normalization, for example, a process of extending the brightness range of the skin image may be performed so that the maximum brightness value of the subject's skin image becomes a predetermined upper limit value. In addition, the brightness value histogram of the subject may be moved so that the average brightness value of the skin image of the subject becomes a predetermined value.

  Part or all of the above contents can also be specified as follows. However, the above-mentioned content is not limited to the following description.

<1> A skin condition evaluation method including an acquisition step of acquiring a blur image having a predetermined blur intensity obtained by photographing a skin surface, and a generation step of generating distribution information of a specific brightness range in the blur image.
<2> The distribution information includes an index value indicating at least one of an area, the number of local regions, and a direction based on pixels in the specific brightness range in the blurred image, or the local region in the blurred image. The skin condition evaluation method according to <1>, which is an image to be represented.
<3> In the acquisition step, a plurality of blurred images with different blur strengths are obtained on the skin surface, and in the generation step, the distribution information is generated for each blur image with different blur strengths. <1> or the skin condition evaluation method according to <2>.
<4> The skin according to <3>, wherein the generation step further generates information indicating a change tendency due to the difference in the blur intensity in the plurality of pieces of distribution information generated for each of the blur images having different blur intensity. State evaluation method.
<5> A step of evaluating the state of the test skin by comparing the information indicating the change tendency generated with respect to the test skin and the information indicating the change tendency generated with respect to the sample skin is further included <4>. The skin condition evaluation method described in 1.
<6> In the obtaining step, the plurality of blurred images having different blurring strengths are obtained by applying a smoothing filter a plurality of times to the skin image obtained by photographing the skin surface. <3> to <5> The skin condition evaluation method as described in any one of these.
<7> The specific brightness range is a range of brightness values selected from a predetermined brightness value range excluding a predetermined range including an upper limit value of brightness that can be indicated by each pixel of the blurred image. <1 > To <6> The skin condition evaluation method according to any one of <6>.
<8> Any one of <1> to <7>, further including a step of normalizing the distribution information based on the distribution information of the specific brightness range extracted from the unblurred skin image The skin condition evaluation method described in 1.
<9> Skin condition before and after application of cosmetics, beauty treatment (excluding medical practice) or cosmetic or health food / drink according to the skin condition evaluation method according to any one of <1> to <8> A method of evaluating the effect of the cosmetic, the cosmetic treatment, or the food or drink by evaluating.
<10> A skin condition evaluation apparatus comprising: an acquisition unit that acquires a blurred image having a predetermined blur intensity obtained by photographing a skin surface; and a generation unit that generates distribution information of a specific brightness range in the blurred image.
<11> The generation unit may use, as the distribution information, an index value indicating at least one of an area, the number of local regions, and a direction based on pixels in the specific brightness range in the blurred image, or in the blurred image The skin condition evaluation apparatus according to <10>, wherein an image representing the local region is generated.
<12> The acquisition unit includes acquiring a skin image obtained by photographing a skin surface, and applying a smoothing filter to the acquired skin image a plurality of times, and a plurality of blurred images having different blur strengths The skin condition evaluation apparatus according to <10> or <11>, wherein the generation unit generates the distribution information for each blurred image having a different blur intensity.
<13> The skin according to <12>, wherein the generation unit further generates information indicating a change tendency associated with a difference in the blur intensity in a plurality of pieces of distribution information generated for each blur image having different blur intensity. Condition evaluation device.
<14> The specific brightness range is a range of brightness values selected from a predetermined brightness value range excluding a predetermined range including an upper limit value of brightness that can be indicated by each pixel of the blurred image. > To <13> The skin condition evaluation device according to any one of <13>.
<15> Any one of <10> to <14>, further comprising normalization means for normalizing the distribution information based on the distribution information of the specific brightness range generated from the skin image without blurring The skin condition evaluation apparatus as described in one.

10 Skin condition evaluation device (evaluation device)
11 CPU
12 Memory 13 Input / output I / F
14 Communication Unit 15 Display Device 16 Input Device 21 Image Acquisition Unit 22 Blur Processing Unit 23 Generation Unit 24 Information Storage Unit 25 Output Processing Unit

Claims (11)

An acquisition step of acquiring a blur image of a predetermined blur intensity obtained by photographing the skin surface;
A generation step of generating distribution information of a specific brightness range in the blurred image;
Skin condition evaluation method including The distribution information is an index value indicating at least one of an area, the number of local regions, and a direction based on pixels of the specific brightness range in the blurred image, or an image representing the local region in the blurred image. is there,
The skin condition evaluation method according to claim 1. In the acquisition step, a plurality of blurred images with different blur strengths, which are blurred images of the skin surface, are acquired,
In the generation step, the distribution information is generated for each blurred image having a different blur intensity.
The skin condition evaluation method according to claim 1 or 2. In the generating step, further generating information indicating a change tendency due to the difference in the blur intensity in a plurality of distribution information respectively generated for each blur image having a different blur intensity.
The skin condition evaluation method according to claim 3. A step of evaluating the state of the test skin by comparing the information indicating the change tendency generated with respect to the test skin and the information indicating the change tendency generated with respect to the sample skin;
The skin condition evaluation method according to claim 4, further comprising: In the obtaining step, the plurality of blurred images having different blur strengths are obtained by applying a smoothing filter multiple times to the skin image obtained by photographing the skin surface.
The skin condition evaluation method according to any one of claims 3 to 5. The specific brightness range is a range of brightness values selected from a predetermined brightness value range excluding a predetermined range including an upper limit value of brightness that can be indicated by each pixel of the blurred image.
The skin condition evaluation method according to any one of claims 1 to 6. Normalizing the distribution information based on the distribution information of the specific brightness range extracted from an unblurred skin image related to the skin surface;
The skin condition evaluation method according to any one of claims 1 to 7, further comprising:   By evaluating the skin before and after application of cosmetics, beauty treatment (excluding medical practice) or cosmetic or health food and drink by the skin condition evaluation method according to any one of claims 1 to 8, A method for evaluating the effect of the cosmetic, the cosmetic treatment or the food or drink. An acquisition means for acquiring a blurred image of a predetermined blur intensity obtained by photographing the skin surface;
Generating means for generating distribution information of a specific brightness range in the blurred image;
A skin condition evaluation apparatus comprising: The acquisition means includes
Get a skin image of the skin surface,
Applying a smoothing filter multiple times to the acquired skin image;
To obtain multiple blurred images with different blur intensities,
The generation unit generates the distribution information for each blurred image having a different blur intensity.
The skin condition evaluation apparatus according to claim 10.