JP6124295B2 - Method for evaluating the texture of bare skin or makeup skin - Google Patents

Description

  The present invention relates to a method for evaluating the visual texture of bare skin or makeup skin.

  In order to evaluate the effect of skin care cosmetics and makeup cosmetics on the visual texture of the skin (hereinafter referred to as “texture”), an optical index value is obtained from an image of bare skin or makeup skin. It was. For example, dispersion or distortion is used as a texture index value from an image of bare skin or makeup skin (Patent Document 1). Further, the texture is caused by unevenness of brightness and color on the surface, and the kind of texture varies greatly depending on what kind of spatial frequency non-uniformity is included. For this reason, a method has been used in which a specular reflection component is extracted from an image of bare skin and an apparent roughness is obtained based on a specific spatial frequency component obtained from the specular reflection component (Patent Document 2).

JP 2009-131336 A JP 2004-166801 A

  The relationship between the texture and the optical index value is not completely clarified. Even in the texture evaluation method using the optical index value, the target texture is evaluated more appropriately. Development of an evaluation method that can be obtained has been desired. Moreover, although unevenness in the skin is often regarded as an element that impairs beauty as “unevenness”, the skin is not originally a uniform surface, and the uniformity causes artificial unnaturalness. Therefore, the unevenness of the low frequency band that the skin should have for beauty, the non-uniformity of the high frequency band that the skin should have for beauty, and the nonuniformity that impairs the beauty. Each needed to be classified appropriately. Further, in the conventional texture evaluation method, it has been desired to facilitate, reduce costs, save resources, and the like.

  The present invention has been made in view of such circumstances, and an object thereof is to quantitatively evaluate the visual texture of bare skin or makeup skin using optical index values. More specifically, it is to appropriately classify the non-uniformity of bare skin or makeup skin, and to quantitatively evaluate the texture resulting from the relationship of each classified non-uniformity.

  In order to solve such a problem, the present inventors acquired spatial frequency components from images obtained by imaging a plurality of cheeks, and calculated the amount of change in each spatial frequency component. It was found that the appearance of “translucency” appears due to the low. Furthermore, from this knowledge, skin non-uniformity is “shading that does not contribute to the texture”, “non-uniformity in the low frequency band that causes translucency”, “nonuniformity that impairs translucency (unevenness)”, “translucency” It has been found that it can be classified into four bands of “non-uniformity (texture) of the high frequency band causing a feeling”, and the present invention has been completed. Hereinafter, the appearance of “translucency” is also simply referred to as “transparency”.

  SUMMARY An advantage of some aspects of the invention is to solve at least a part of the problems described above, and the invention can be implemented as the following forms.

  (1) According to one aspect of the present invention, a method for evaluating the texture of bare skin or makeup skin is provided. The method includes (a) obtaining a captured image of bare skin or makeup skin, (b) obtaining a component in each spatial frequency band of the captured image, and (c) at least four components in each spatial frequency band. A step of classifying into a plurality of bands equal to or higher than a band; (d) a step of obtaining an index value of the change amount for each of the plurality of bands; And (f) evaluating the texture of bare skin or makeup skin based on the texture index value. According to the method of this aspect, the captured image of bare skin or makeup skin is classified for each spatial frequency component and an index value is obtained, so that the texture generated from the index value of each band can be objectively evaluated. In other words, it is possible to objectively evaluate the change in texture caused by the change in the index value of each band that occurs when using skin care cosmetics or makeup cosmetics. The “texture resulting from the relationship between the change amounts of the plurality of bands” is, for example, from the relationship between the change amounts of the second to fourth bands when classified into the four bands of the first to fourth bands. There is a sense of see-through caused by a relatively low amount of change. The classification of the step (c) is more preferably classified into four bands, but can be further subdivided within a range not impairing the gist of the present invention.

  (2) In the method of the above aspect, the classification of the step (c) is a band that can be seen through by a psychophysical experiment using an image from which each spatial frequency component is removed (hereinafter referred to as “the band”). ) Are extracted and classified as four bands: first band: 1 cycle / image width, second band: 2 cycle / image width to less than the band, third band: the band, and fourth band: a band higher than the band. May be. According to this form of the method, classification is performed according to the human texture perception mechanism, and therefore, classification can be performed appropriately.

  (3) In the method of the above aspect, the classification in the step (c) is based on the width of the facial cheek part as follows: first band: 1 cycle / cheek width, second band: 2-8 cycles / cheek width, third band : 16 to 32 cycles / cheek width, fourth band: 64 cycles ~ / cheek width may be classified as four bands. According to this method, the analysis result can be obtained in a short time.

(4) In the method of the above aspect, in the step (e), the index values of the change amounts of the _second to _fourth bands when classified into the four bands of the first to fourth bands are expressed as SD ₂ to ₄ , respectively. A value obtained by 1 may be used as an index value. In Expression 1, a negative value is taken when the index value of the change amount of the third band is larger than the index value of the change amount of the second or fourth band, and a higher value is taken as it is relatively smaller. According to the method of this embodiment, it is possible to evaluate a sense of sheer caused by a relatively low change amount of the third band in a short time.
(Expression 1) Index value = (SD ₂ −SD ₃ ) × (SD ₄ −SD ₃ )
However, if SD ₃ > SD ₂ and SD ₃ > SD ₄ , the evaluation is impossible.

  (5) In the method of the above aspect, in the step (e), the index value of the change amount of the second band when the four bands of the first to fourth bands are classified, It may be an index value of strength. According to this form of the method, it is possible to evaluate in a short time the strength of non-uniformity that the skin classified according to the human texture perception mechanism should have.

  (6) In the method of the above aspect, in the step (e), the index value of the change amount of the third band when classified into the four bands of the first to fourth bands is used as the index value of the intensity of unevenness. Also good. According to this form of the method, the intensity of unevenness classified according to the human texture perception mechanism can be evaluated in a short time.

  (7) In the method of the above aspect, in the step (e), the index value of the change amount of the fourth band when the four bands of the first to fourth bands are classified is used as the index value of the strength of the texture. Also good. According to this form of the method, the strength of the texture classified according to the human texture perception mechanism can be evaluated in a short time.

  (8) In the method of the above aspect, the step (b) may use wavelet transform. According to the method of this embodiment, each spatial frequency component can be easily obtained, so that an analysis result can be obtained in a short time.

  (9) In the method of the above aspect, the index value in the step (d) may be a value statistically obtained from the pixel value. According to the method of this embodiment, since the value obtained by the statistical method is used as the index value, an objective analysis result can be obtained.

  The present invention can be realized in various modes. For example, the present invention can be realized in the form of an apparatus that implements at least a part of the above-described evaluation method, a computer program, or a non-temporary recording medium that records the computer program.

It is a flowchart which shows the procedure of the evaluation method of the texture in this embodiment. It is a flowchart which shows the procedure of the method which classify | categorizes a spatial frequency component into 4 bands in step S115. It is explanatory drawing which shows the process result of step S210 typically. It is explanatory drawing which shows an example of the parameter | index value of the texture in step S125. It is explanatory drawing which shows the captured image of 1st Example.

A. Embodiment A1. Bare or Makeup Skin Texture Evaluation Method FIG. 1 is a flowchart showing a procedure of a bare skin or makeup skin texture evaluation method in the present embodiment. In the present embodiment, the makeup skin means skin applied with any one or a combination of any cosmetics that can be applied to the skin, such as a makeup base, foundation, funnier, and teak.

  As shown in FIG. 1, first, a captured image of bare skin or makeup skin is obtained (step S105). As the imaging apparatus, any device that can obtain digital image data, such as a digital still camera or a digital video camera, can be employed. However, it is possible to obtain an image having a resolution sufficient for analyzing a component (for example, texture) that contributes to the texture of the cheek and an imaging range sufficient for analyzing a spatial change in luminance and color of the entire cheek. Possible devices are preferred. The captured image is preferably trimmed to fit the cheek width. In the following description, luminance is used as the pixel value. However, not only the luminance but also pixel values such as hue and saturation can be adopted.

The captured image obtained in step S105 is decomposed into spatial frequency components (step S110). As a specific processing procedure, first, luminance value data of each pixel is acquired from a captured image, and each spatial frequency component is obtained by wavelet transform or the like. By using an image obtained by trimming the captured image according to the cheek width, a spatial frequency component of 2 ⁿ cycles / cheek width (n = 0, 1, 2,...) Can be obtained with the cheek width as a reference. . When the wavelet transform is used, Haar's basis, which is the simplest algorithm, can be adopted, and it can be executed by creating a program and mounting it on a computing device such as a personal computer.

  The obtained spatial frequency components are classified into four bands of the first to fourth bands (step S115). In the present embodiment, as a classification method, with reference to the cheek width of the face, the first band: 1 cycle / cheek width, the second band: 2-8 cycles / cheek width, the third band: 16-32 cycles / cheek width, the fourth Band: adopts a method of classifying as 64 cycle ~ / cheek width.

  Each band component obtained in step S115 is integrated, and an index value of the change amount of each band is acquired (step S120). The integration can be executed by performing an addition process using an arithmetic device such as a personal computer. In addition, as the index value, an index value representing the amount of change such as variance or standard deviation can be adopted.

  From the index value of the change amount of each band component obtained in step S120, the texture index value generated by the relationship is acquired (step S125). In this step, a calculation method can be employed in which the above-described Equation 1 is used as the index value of the sense of transparency based on the index value.

  As the texture index value, in addition to the above-described index value of translucency, the index value of the change amount of each band can be directly adopted as the texture index value. Specifically, the index value of the amount of change in the second band is used as an index value of the intensity of non-uniformity that the skin should have, and the index value of the amount of change in the third band is used as an index value of skin unevenness. It is possible to adopt a method that uses an index value for the amount of change in the fourth band as an index value for skin texture. Note that the first band includes shadow information and has a low contribution to the texture.

  When the index value is calculated, the texture of the captured skin or makeup skin is evaluated based on the obtained index value (step S130). Specifically, among the above-described index values of the texture, when the index value of the sense of sheer is adopted, it is evaluated that the sense of sheer is higher as the index value is higher.

  Since the texture of bare skin or makeup skin can be evaluated as described above, the texture control effect can be taken into account when selecting cosmetics. For example, when selecting a basic cosmetic product that has an effect of improving the texture of the skin, it is possible to grasp the texture of the skin before and after continuous use by performing the above-described processing using the images taken before and after the continuous use, resulting in a high texture improvement effect. Based on this, it is possible to determine which basic cosmetic product is selected. As another example, when selecting a makeup cosmetic, a plurality of makeup cosmetics are manufactured, and by performing the above-described processing for these samples, the texture of the skin made by each sample is grasped, Based on the effect of controlling the texture, it can be determined which makeup cosmetic is selected. Note that the amount of change of each band component can be controlled by various factors such as the makeup cosmetic form, the raw materials to be blended in the makeup cosmetic, the form and material of the applicator, and the combination method of a plurality of makeup cosmetics. Therefore, the above-mentioned elements can be adjusted based on the evaluation result of the texture control effect.

A2. Band Classification Method FIG. 2 is a flowchart showing a processing procedure for classifying spatial frequency components into four bands in step S115 described above. As described above, in step S115 of the present embodiment, the classification of each spatial frequency component is based on the cheek width of the face as a reference: first band: 1 cycle / cheek width, second band: 2-8 cycles / cheek width, third The band is classified as 16 to 32 cycles / cheek width, and the fourth band: 64 cycles to / cheek width. Such classification can be determined as follows.

  First, a noise image having a constant change amount for each spatial frequency component is created (step S205). The noise image is created by, for example, decomposing a grayscale white noise image into each spatial frequency component by wavelet transform, and setting the standard deviation on a scale of 0.0 to 1.0 as 0. It is created by standardizing to 1 and integrating those spatial frequency components again. As the size of the image, for example, an image having 256 pixels in both vertical and horizontal directions can be adopted. Further, it is desirable to create a plurality of types of noise images with different phases.

  Based on the noise image obtained in step S205, an image group from which only a specific band component is removed is created (step S210). For example, an image from which only a component of 1 cycle / image width is removed or an image from which only a component of 32 cycles / image width is removed is created. As a method of creating an image from which only a specific band component is removed, for example, it can be created by decomposing a noise image into each spatial frequency component by wavelet transform and extracting and integrating only the specific band component.

  FIG. 3 is an explanatory diagram showing the image created in step S210. Each image is an image created by removing the spatial frequency component shown below each image from noise with a constant change amount for each spatial frequency component.

  When the image group from which the specific band component is removed is obtained, the band is specified by the subject through a psychophysical experiment (step S215). As shown in FIG. 3, in some images from which a specific spatial frequency component is removed, a plurality of different layers exist in the foreground and the back rather than a single layer so that the foreground layer is transparent. Perceived.

  In the example shown in FIG. 4, after four subjects have adapted to a dark room for 10 minutes, images are presented in a random order at a distance of 60 cm from a 72 dpi presentation device. A psychophysical experiment was performed by selecting from the three answers of “point)”, “slightly feel (score 0.5 points)”, and “feel (score 1.0 points)”. The answers were made six times with the noise phase changed for the same image. It should be noted that “transparent” here is defined more strictly as being perceived as overlapping because different layers are transmitting, and was explained to the subject as well. All subjects were either naked or corrected and had normal visual acuity.

  As shown in FIG. 4, the sense of see-through is amplified in an image from which band components of 16 cycles / image width and 32 cycles / image width are removed. This means that a sense of see-through occurs by removing components of 16 cycles / image width and 32 cycles / image width. Therefore, according to the example of FIG. 4, 16 to 32 cycles / image width can be determined as a transparent band by removing in step S210. It should be noted that the sense of transparency is reduced by removing the band of 2 to 8 cycles / image width according to the example of FIG. Therefore, it can be seen that the band of 2 to 8 cycles / image width is a band necessary for producing a sense of transparency, and the band of 16 to 32 cycles / image width should be classified.

  When the transparent band is specified by extracting, the classification of each spatial frequency component is determined (step S220). Specifically, the first band: 1 cycle / image width, the second band: 2 cycle / image width to less than the band, the third band: the band, and the fourth band: a band higher than the band. Therefore, according to the example of FIG. 4, the first band: 1 cycle / image width, the second band: 2-8 cycles / image width, the third band: 16-32 cycles / image width, and the fourth band: 64 cycles / image width. Can be classified. Based on the above classification, when the index value of the translucency obtained by the above-described equation 1 is calculated from the image group obtained in step S210, it is high in an image from which band components of 16 cycles / image width and 32 cycles / image width are removed. It takes a value and takes a value close to 0 or a negative value in other images.

  According to the texture evaluation method in the embodiment described above, the classification of spatial frequency components is determined by psychophysical experiment, and therefore, appropriate classification according to the human texture perception mechanism can be performed. Therefore, by obtaining an optical index value based on appropriately classified spatial frequency components, "non-uniformity in the low frequency band that causes a sense of sheer", "non-uniformity that impairs the sheer feeling (unevenness)", It is possible to objectively classify and evaluate “non-uniformity (texture) in the high-frequency band causing a sense of see-through” according to the human texture perception mechanism.

  Furthermore, it is possible to quantitatively evaluate the texture that has not been clarified heretofore and that arises from the relationship between a plurality of spatial frequency components. Specifically, according to the texture evaluation method in the embodiment of the present invention, a band that can be seen through is specified by a psychophysical experiment. In other words, the third band that is transparent because there is no change amount relative to the other bands and the second band that is a band that requires a change amount in order to generate a sense of see-through are specified. From the relationship between the change amounts of the second to fourth bands, which are index values, it is possible to quantitatively evaluate the sense of transparency caused by the relative low change amount of the third band.

  Also, wavelet transform is performed on the basis of the image of the cheek image to acquire each spatial frequency component, and the optical index value of the skin texture is obtained on the basis thereof, so that the analysis result can be obtained in a short period of time. .

B. Example In the example, in the above-described step S105, a one-month continuous test with a cosmetic liquid that enhances the permeability of the skin stratum corneum is performed, and each cheek is strobed by a digital camera before the continuous test is started and after the one-month continuous test. Images were taken under light and trimmed to fit the cheek width.

  FIG. 5 is an explanatory diagram using images before and after the continuous test of the example. FIG. 5A shows an image before the continuous test is started. FIG. 5B shows an image after a one-month continuous test. In addition, this image was visually selected by a plurality of experts as an image in which the effect of enhancing the translucency of the skin was most apparent among images obtained by photographing nine consecutive test subjects before and after continuous use.

  Spatial frequency components are acquired for each of the two captured images by wavelet transform (step S110), and each spatial frequency component is determined for the first band: 1 cycle / cheek width, the second band: 2-8 cycles / cheek width, and third. Bands were classified into four bands as 16 to 32 cycles / cheek width and fourth band: 64 cycles to / cheek width (step S115).

The luminance (L * value) standard deviation of each band acquired in the two images described above is calculated as an index value of the change amount of each band (step S120), and the value obtained by the above-described equation 1 is used as a transparent feeling. The index value was calculated (step S125). As a result, the value before the start of the continuous test was −0.6 × 10 ⁻³ and 1.5 × 10 ⁻² after the continuous test for one month. Therefore, in step S130, it was evaluated that the sense of sheer was higher after one month of continuous testing. Since this image is an image selected as an image that has the most effect of enhancing the sense of transparency, the results obtained by the above-described evaluation method are in accordance with sensory evaluation.

C. Modification C1. Modification 1
In the above-described embodiments and examples, each spatial frequency component is acquired by performing wavelet transform in step S110. However, instead of wavelet transform, each spatial frequency component can also be obtained using Fourier transform. . In other words, in general, any method capable of spatial frequency analysis of a captured image can be employed.

C2. Modification 2
In the above-described embodiments and examples, in step S120, the spatial frequency components classified into the four bands are integrated in order to obtain the index value. However, this is omitted, and the index value of each band is assigned to each band. You may obtain by calculating an average value from the index value of the applicable spatial frequency component.

C3. Modification 3
In the embodiments and examples described above, when the imaging of the skin (step S105) has already been performed separately from the texture evaluation, step S105 in the texture evaluation procedure can be omitted. In this case, a process of “acquiring a skin image” is performed instead of step S105. The acquisition of the skin image means, for example, a process of inputting skin image data to the arithmetic device that executes Steps S110 to S125.

  The present invention is not limited to the above-described embodiments, examples, and modifications, and can be realized with various configurations without departing from the spirit thereof. For example, the technical features in the embodiments and the modifications corresponding to the technical features in each embodiment described in the summary section of the invention are to solve some or all of the above-described problems, or In order to achieve part or all of the effects, replacement or combination can be performed as appropriate. Further, if the technical feature is not described as necessary in the present specification, it can be appropriately deleted.

Claims (8)

A method for evaluating the visual texture of bare skin or makeup skin,
(A) obtaining a captured image of bare skin or makeup skin;
(B) obtaining a component in each spatial frequency band of the captured image;
(C) Extracting a band (the band) that can be seen through by extracting the component in each spatial frequency band by a psychophysical experiment using an image from which each spatial frequency component is removed, and the first band: 1 cycle / image Classifying into four bands of width, second band: 2 cycle / image width to less than the band, third band: the band, and fourth band: a band higher than the band ;
(D) obtaining an index value of the change amount of the four bands ;
(E) obtaining a visual texture index value generated from the index value of the change amount of the four bands ;
(F) evaluating the visual texture of bare skin or makeup skin based on the visual texture index value;
A method comprising: The method of claim 1 , wherein
The classification of the step (c) is based on the width of the facial cheek part, the first band: 1 cycle / cheek width, the second band: 2-8 cycles / cheek width, the third band: 16-32 cycles / cheek width, 4 bands: A method of classifying as 4 bands of 64 cycles ~ / cheek width. The method according to claim 1 or 2 , wherein
The step (e) is a method in which the index value of the amount of change in the second to fourth bands is SD _{2 to 4} , respectively, and the value obtained by Equation 1 is used as the index value of the sense of transparency.
(Expression 1) Index value = (SD ₂ −SD ₃ ) × (SD ₄ −SD ₃ )
However, if SD ₃ > SD ₂ and SD ₃ > SD ₄ , the evaluation is impossible. The method according to claim 1 or 2 , wherein the step (e) uses the index value of the amount of change in the second band as an index value of the strength of non-uniformity that the skin should have. 3. The method according to claim 1 , wherein the step (e) uses an index value of the amount of change in the third band as an index value of unevenness intensity. 3. The method according to claim 1 , wherein the step (e) uses an index value of the amount of change in the fourth band as an index value of texture strength. In the method in any one of Claims 1-6 ,
The method wherein step (b) is wavelet transform. In the method in any one of Claims 1-7 ,
The method, wherein the index value of the change amount in the step (d) is a value statistically obtained from the pixel value.