Method for determining UV protection efficacy of functional cosmetics on basis of ai, and system therefor

Abstract

The present invention relates to a method for determining UV protection efficacy of functional cosmetics on the basis of AI and, particularly, to a UV protection efficacy determination method comprising: applying functional cosmetics to a subject's skin; subsequently irradiating the skin with UV light; and determining protection efficacy by observing changes in the skin, wherein the observing of the changes in the skin irradiated with UV light is performed through detection of erythema or pigment darkening by an artificial intelligence algorithm from an image obtained by capturing the subject's skin; and the artificial intelligence algorithm performs the detection of erythema or pigment darkening by applying skin tone information calculated from an image obtained by capturing the subject's skin not irradiated with UV light. According to the present invention, UV protection efficacy is determined using artificial intelligence, so that the determination of the UV protection efficacy is objectified, and the reliability of determination results increases. In addition, since the determination process is automated by artificial intelligence, time required for determination can be reduced.

Description

AI-based method and system for determining UV protection efficacy of functional cosmetics

[0001] The present invention relates to a method and system for determining the UV blocking efficacy of functional cosmetics, and more specifically, to a method for determining the UV blocking efficacy of functional cosmetics using artificial intelligence and a system applying the same.

[0002] Generally, ultraviolet (UV) radiation from the sun acts as a major cause of skin conditions such as erythema, edema, freckles, and skin cancer. As these symptoms pose problems in both health and cosmetic aspects, research on UV-induced skin diseases is actively underway. Furthermore, the Korea Meteorological Administration publishes UV indices to encourage the public to protect themselves from UV rays, reflecting a growing trend to minimize public health issues related to UV radiation, such as skin cancer and cataracts.

[0003] These ultraviolet rays are classified according to wavelength into ultraviolet C (hereinafter "UV-C") with a wavelength of 200-290 nm, ultraviolet B (hereinafter "UV-B") with a wavelength of 290-320 nm, and ultraviolet A (hereinafter "UV-A") with a wavelength of 320-400 nm. UV-C passes through the ozone layer and is lost without reaching the Earth's surface, while UV-B penetrates to the epidermis of the skin and causes erythema, freckles, and edema. In addition, UV-A is known to penetrate to the dermis of the skin and cause skin aging and skin irritation, such as promoting skin cancer, wrinkles, and melanin formation.

[0004] Recently, many functional cosmetics with UV-blocking capabilities designed to protect the skin from ultraviolet rays are being sold, and the government has enacted administrative rules to enable users to easily identify the UV-blocking performance of these functional cosmetics. The Ministry of Food and Drug Safety Notice regarding "Methods and Standards for Measuring UV-blocking Effects" establishes the measurement methods and standards for the UV-A protection factor and UV-A blocking effect of functional cosmetics that help to tan the skin evenly or protect it from ultraviolet rays, in accordance with Article 9 of the Cosmetics Act.

[0005] In accordance with the "Methods and Standards for Measuring Ultraviolet Protection Effects," the "Sun Protection Factor (SPF)" or the "Protection grade of UVA," abbreviated as PA, can be evaluated.

[0006] In the "Methods and Standards for Measuring Ultraviolet Protection Effect," the minimum erythema dose (MED) is applied as an indicator of the ultraviolet protection factor (SPF), and the minimal persistent pigment darkening dose (MPPD) is applied as an indicator of the ultraviolet A protection rating (PA).

[0007] However, since the assessment of such UV blocking efficacy is determined visually by two or more skilled individuals under a sufficiently bright light source, the results may be subjective and there is a disadvantage in that it is difficult to accurately reflect subtle changes.

[0008] Meanwhile, artificial intelligence is a subfield of computer science that seeks to artificially replicate human learning, reasoning, and perceptual abilities, and deep learning technology, a machine learning algorithm capable of high levels of abstraction, is receiving significant attention.

[0009] In particular, deep learning technology is actively researching image recognition technology that classifies specific regions or objects appearing in an image into specific categories (class), and such image classification-based image recognition technology is based on a deep neural network structure.

[0010] Although technology is being developed to detect and analyze various skin symptoms in images of facial skin using such deep learning technology, it is not being applied as a technology to determine UV protection efficacy.

[0011]

[0012] The present invention aims to solve the problems of the aforementioned prior art by providing a method for determining the UV blocking efficacy of functional cosmetics using artificial intelligence.

[0013] The method for determining the UV blocking efficacy of an AI-based functional cosmetic according to the present invention for achieving the above objective is a method for determining UV blocking efficacy by applying a functional cosmetic to a subject's skin, irradiating it with ultraviolet rays, and observing skin changes to determine the blocking efficacy, wherein the process of observing skin changes irradiated with ultraviolet rays is performed by detecting erythema or darkening using an artificial intelligence algorithm on an image of the subject's skin, and the artificial intelligence algorithm is characterized by performing erythema or darkening detection by reflecting skin tone information derived from an image of the subject's skin not irradiated with ultraviolet rays.

[0014] The above skin tone information may be obtained by calculating the average RGB values ​​of an image of a subject's skin that has not been irradiated with ultraviolet rays, converting it into L*, a*, and b* color spaces, and calculating the Individual Typology Angle (ITA) for the angle formed by the L and b axes.

[0015] The above skin tone information may be obtained by converting RGB into L*, u*, v* color spaces and calculating a UV histogram from an image of a subject's skin that has not been irradiated with ultraviolet rays.

[0016] The above skin tone information may be the average RGB value of an image of a subject's skin that has not been irradiated with ultraviolet rays.

[0017] According to another embodiment of the present invention, an AI-based system for determining the UV blocking efficacy of a functional cosmetic is a system for determining UV blocking efficacy applied in a process of determining the blocking efficacy by observing skin changes after applying a functional cosmetic to the skin of a subject and irradiating it with ultraviolet rays. The system comprises: an image acquisition unit for acquiring an image of the subject's skin; a skin tone information calculation unit for calculating information regarding skin tone from an image of the skin not irradiated with ultraviolet rays among the images of the subject's skin; and a symptom detection unit for detecting erythema or melanosis in an area irradiated with ultraviolet rays among the images of the subject's skin. The symptom detection unit detects erythema or melanosis in an area irradiated with ultraviolet rays by an artificial intelligence algorithm, and the artificial intelligence algorithm detects erythema or melanosis by reflecting the information regarding skin tone calculated by the skin tone information calculation unit.

[0018] The above skin tone information calculation unit may calculate the average RGB value of an image of the skin of a subject not irradiated with ultraviolet rays, convert it into an L*, a*, b* color space, and calculate an ITA (Individual Typology Angle) for the angle formed by the L and b axes.

[0019] The above skin tone information calculation unit may convert RGB into L*, u*, and v* color spaces in an image of the skin of a subject not irradiated with ultraviolet rays, and calculate a UV histogram.

[0020] The above skin tone information calculation unit may calculate the average RGB value of an image of a subject's skin that has not been irradiated with ultraviolet rays.

[0021]

[0022] The present invention, configured as described above, determines the UV blocking efficacy using artificial intelligence, thereby objectifying the determination of UV blocking efficacy and increasing the reliability of the determination result.

[0023] In addition, the automation of the judgment process by artificial intelligence has the effect of reducing the time required for judgment.

[0024]

[0025] Figure 1 is a diagram showing the separation of a skin area not irradiated with ultraviolet rays from an image taken to include an area irradiated with ultraviolet rays.

[0026] Figure 2 is an example of classifying skin types based on skin color using ITA.

[0027] Figure 3 is an example of a UV histogram transformed image converted from an RGB image.

[0028] Figure 4 is a diagram illustrating the detection of erythema or blackening areas by an artificial intelligence algorithm in an image captured to include the ultraviolet irradiation area.

[0029] FIG. 5 is a schematic diagram illustrating the configuration of an AI-based system for determining the UV protection efficacy of functional cosmetics according to an embodiment of the present invention.

[0030]

[0031] An embodiment according to the present invention will be described in detail with reference to the attached drawings.

[0032] However, embodiments of the present invention may be modified in various other forms, and the scope of the present invention is not limited only to the embodiments described below. The shapes and sizes of elements in the drawings may be exaggerated for clearer explanation, and elements indicated by the same reference numerals in the drawings are the same elements.

[0033] Furthermore, throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected" but also cases where they are "electrically connected" with other components interposed between them. Additionally, when a part is described as "including" or "equipped" with a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but rather allows for the inclusion or equipping of additional components.

[0034] Furthermore, terms such as "first," "second," etc., are intended to distinguish one component from another, and the scope of rights shall not be limited by these terms. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0035]

[0036] The present invention relates to a method for determining the UV-blocking efficacy of a functional cosmetic having a UV-blocking effect, characterized by applying artificial intelligence rather than the subjective judgment of an examiner in the process of determining the blocking efficacy by observing skin changes after applying the functional cosmetic to the skin of a subject and irradiating it with UV rays.

[0037] Therefore, excluding the part where the examiner's judgment is involved in the process of determining UV protection efficacy, the existing process for determining UV protection efficacy can be applied as is, and various criteria for determining UV protection efficacy and the corresponding inspection processes can all be applied within a scope that does not impair the features of the present invention.

[0038] The present invention is described below by reflecting the process of measuring the ultraviolet blocking effect in accordance with the "Method and Standards for Measuring Ultraviolet Blocking Effect" (Korea Food and Drug Administration Notification No. 2012-88), but is not limited thereto.

[0039]

[0040] "Method and Standard for Measuring UV Protection Effect" selects at least 10 subjects per product and performs measurements on the subjects' backs.

[0041] To measure the Sun Protection Factor (SPF), the minimum erythema dose is measured on the area without product application and the area with product application, and to measure the Protection Factor A (PFA), the minimum sustained immediate darkening dose is measured on the area without product application and the area with product application.

[0042] At this time, the product application area is 24cm 2 0.5cm 2 Five or more inspection areas having the above area are partitioned.

[0043] And the amount of light at each irradiation site is adjusted so that the site where minimal erythema or minimally sustained immediate blackening is expected becomes the middle (e.g., position 3 or 4), and the amount of light is increased at equal intervals accordingly.

[0044] Finally, the Sun Protection Factor (SPF) is determined by calculating the Minimum Erythema Dose (MEDu) of the unapplied area and the Minimum Erythema Dose (MEDp) of the product-applied area, calculating the SPF for each subject according to a predetermined formula, and determining the arithmetic mean value; and the Ultraviolet A Protection Factor (PFA) is determined by calculating the Minimum Long-lasting Immediate Darkening Dose (MPPDu) of the unapplied area and the Minimum Long-lasting Immediate Darkening Dose (MPPDp) of the product-applied area, calculating the PFA for each subject according to a predetermined formula, and determining the arithmetic mean value.

[0045] Among the above-mentioned processes, the measurement of the minimum erythema dose in the area not covered by the product, the measurement of the minimum erythema dose in the area covered by the product, the measurement of the minimum sustained immediate darkening dose in the area not covered by the product, and the measurement of the minimum sustained immediate darkening dose in the area covered by the product are performed by two or more skilled inspectors. The criterion for determining the minimum erythema dose is the lowest amount of ultraviolet radiation applied to the covered areas where erythema of 50% or more was generated among multiple irradiated areas, and the criterion for determining the minimum sustained immediate darkening dose is the lowest amount of ultraviolet radiation applied to the covered areas where darkening of 50% or more occurred among multiple irradiated areas.

[0046] The method for determining the UV protection efficacy of an AI-based functional cosmetic according to the present embodiment is characterized by the fact that the process of measuring the minimum erythema dose of the area not covered by the product, measuring the minimum erythema dose of the area covered by the product, measuring the minimum sustained-release immediate darkening dose of the area not covered by the product, and measuring the minimum sustained-release immediate darkening dose of the area covered by the product is performed by artificial intelligence.

[0047] However, when using artificial intelligence to detect erythema and darkening based on the same criteria to determine the light intensity corresponding to the minimum erythema dose or the minimum sustained immediate darkening dose, a problem arises in which errors occur in the measurement results because characteristics based on differences in the subject's skin tone are not reflected. In this case, it is stated that the difference in skin tone is reflected by taking into account the influence of ultraviolet rays, unlike general skin color.

[0048] For example, people with darker skin tones generally have a relatively higher amount of melanin, which absorbs ultraviolet rays and protects skin cells; consequently, even when exposed to the same light source, the degree of erythema or darkening appears weaker. Therefore, if these skin tone characteristics are not taken into account, the UV protection effect of functional cosmetics may be evaluated differently from reality.

[0049] Conventionally, examiners reflected characteristics based on skin tone, but the degree of reflection varied depending on the examiner, and there was a problem in that it was difficult to establish objective standards. In addition, conventionally, even in the process of performing judgment by artificial intelligence, the judgment results were inaccurate because the subject's skin tone could not be reflected.

[0050]

[0051] This embodiment is characterized by applying artificial intelligence in the process of confirming skin erythema or darkening, while simultaneously reflecting the difference in the subject's skin tone.

[0052] To this end, in this embodiment, a domain conversion is performed to convert image information of skin not irradiated with ultraviolet rays into skin tone information.

[0053] For image information regarding skin not irradiated with ultraviolet rays, a separately captured image of the skin may be used, but it may also be applied by excluding the irradiated area from an image captured to include the irradiated area.

[0054] Figure 1 is a diagram showing the separation of a skin area not irradiated with ultraviolet rays from an image taken to include an area irradiated with ultraviolet rays.

[0055] One or more of the following three methods can be applied to convert into skin tone information.

[0056] First, a transformation using ITA (Individual Typology Angle) can be applied.

[0057] The average RGB values ​​are calculated from images of skin not irradiated with ultraviolet rays, converted into L*, a*, and b* color spaces, and the ITA characterizing the skin color parameters is used, with the following calculation formula applied.

[0058]

[0059] Figure 2 is an example of classifying skin types based on skin color using ITA.

[0060] As described, ITA derives the angle formed by the L and b axes, and divides the derived angle into predetermined ranges to be used as a criterion for classifying skin types based on skin color.

[0061] In this embodiment, the angle of the ITA itself is used as skin tone information. At this time, the ITA-converted skin tone information may be in the form of an angle, or it may be in a normalized form obtained by dividing by the maximum angle value.

[0062] Next, a transformation using UV histograms can be applied.

[0063] In an image of skin not irradiated with ultraviolet rays, RGB is converted to L*, u*, v* color spaces to calculate the UV histogram.

[0064] Figure 3 is an example of a UV histogram transformed image converted from an RGB image.

[0065] The method using UV histograms derives skin tone information by using chromaticity (u) and hue (v) information excluding brightness (L) in the L*, u*, and v* color spaces, and has the advantage of reducing the influence of lighting intensity when deriving information on the examiner's skin tone.

[0066] Finally, the average RGB values ​​can be calculated from images of skin not irradiated with UV rays and reflected as information regarding skin tone.

[0067]

[0068] When applying the determination method of this embodiment to the "Method and Standard for Measuring Ultraviolet Protection Effect," the occurrence of erythema is determined for multiple ultraviolet irradiated areas (irradiated patch areas) in the measurement of the minimum erythema dose on the area without product application and the measurement of the minimum erythema dose on the area with product application, and the occurrence of blackening is determined for multiple ultraviolet irradiated areas (irradiated patch areas) in the measurement of the minimum sustained immediate blackening dose on the area without product application and the measurement of the minimum sustained immediate blackening dose on the area with product application.

[0069] To perform this process using artificial intelligence, the UV-irradiated areas (irradiation patch regions) are first identified in the captured skin image.

[0070] Then, erythema or blackening is detected in the separated UV-irradiated areas, and by comparing the entire area of ​​each UV-irradiated area with the detected erythema or blackening area, the UV-irradiated area in which erythema or blackening appears on the entire surface of the UV-irradiated area is determined and selected among the multiple UV-irradiated areas.

[0071] Figure 4 is a diagram illustrating the detection of erythema or blackening areas by an artificial intelligence algorithm in an image captured to include the ultraviolet irradiation area.

[0072] At this time, the process of distinguishing the UV-irradiated area in the captured skin image and the process of detecting erythema or melanosis in the distinguished UV-irradiated area may utilize various general artificial intelligence algorithm technologies within a scope that does not impair the features of the present invention.

[0073] However, in this embodiment, the previously converted skin tone information of the subject is reflected in the process of detecting erythema or darkening in the UV-irradiated area by artificial intelligence.

[0074] A method for reflecting skin tone information can be applied in accordance with an artificial intelligence algorithm that detects erythema or darkening.

[0075] When erythema or darkening is detected by an artificial intelligence algorithm that reflects skin tone information, it is possible to objectively and accurately determine the UVB and UVA light amounts corresponding to the minimum erythema dose in the product-unapplied area and the minimum sustained immediate darkening dose in the product-applied area, as well as the minimum sustained immediate darkening dose in the product-unapplied area and the product-applied area, regardless of the examiner's skin characteristics.

[0076] In this case, since the AI ​​algorithm detecting erythema or darkening receives skin tone information as input, the model learns autonomously to predict detection results that reflect the influence of skin tone by combining these elements. Therefore, there is no need to create additional labels for skin tone; the model can independently learn the features necessary to reflect skin tone during the training process even if only the judgment result labels are provided.

[0077]

[0078] FIG. 5 is a schematic diagram illustrating the configuration of an AI-based system for determining the UV protection efficacy of functional cosmetics according to an embodiment of the present invention.

[0079] A system for determining the UV blocking efficacy of an AI-based functional cosmetic according to another embodiment of the present invention is configured to perform, by artificial intelligence, a process of determining the blocking efficacy by observing skin changes after applying a functional cosmetic to a subject's skin and irradiating it with ultraviolet rays, and includes an image acquisition unit (100), a skin tone information calculation unit (200), and a symptom detection unit (300).

[0080] The image acquisition unit (100) is a component that acquires an image of the subject's skin, which is the target for detecting skin changes, namely erythema or darkening. The image of the subject's skin may be an image received from another location or an image taken at the site. If an image received from another location is used, the image acquisition unit (100) may be a device that receives image data via wired or wireless transmission or receives image data through a storage device. If an image taken at the site is used, the image acquisition unit (100) may be a camera, and any means capable of capturing an image may be applied without limitation.

[0081] The skin tone information calculation unit (200) is a component that calculates information regarding skin tone from an image of skin that has not been irradiated with ultraviolet rays among images of the subject's skin. The information of the image is converted into information regarding skin tone by an artificial intelligence or image processing process and obtained. Any device capable of processing such a process can be applied without limitation as the skin tone information calculation unit (200), and a processor that processes digital data can be applied.

[0082] As previously explained, the skin tone information calculation unit (200) may apply ITA transformation, apply UV histogram transformation, or use the average RGB value of the image in an image of skin that has not been irradiated with ultraviolet rays.

[0083] The symptom detection unit (300) is a component that detects skin changes, namely erythema or darkening, in the ultraviolet irradiated area (irradiation patch area) where ultraviolet rays are irradiated, among images of the subject's skin. The symptom detection unit (300) is performed by an artificial intelligence algorithm that detects erythema or darkening, and additional processes may be performed according to a specific judgment method. For example, in the "method and standard for measuring ultraviolet blocking effect," among multiple ultraviolet irradiated areas, the area where erythema or darkening appears on the entire surface of the ultraviolet irradiated area must be determined, and the symptom detection unit (300) can determine and select the ultraviolet irradiated area where erythema or darkening appears on the entire surface of the ultraviolet irradiated area among multiple ultraviolet irradiated areas by comparing the total area of ​​each ultraviolet irradiated area with the detected erythema or darkening area.

[0084] The process of detecting erythema or darkening in the area irradiated with ultraviolet rays may utilize various general artificial intelligence algorithm technologies within a scope that does not impair the features of the present invention, provided that the detection of erythema or darkening is characterized by reflecting information regarding the subject's skin tone calculated by the skin tone information calculation unit (200). Since the configuration for applying the subject's skin tone information during the detection process is the same as previously described, a detailed explanation is omitted. A device capable of processing such an artificial intelligence algorithm may be applied without limitation as the symptom detection unit (300), and a processor for processing digital data may be applied.

[0085] When applying the AI-based system for determining the UV protection efficacy of functional cosmetics according to the present embodiment, the process of measuring the minimum erythema dose in the area without product application and the minimum erythema dose in the area with product application, as well as measuring the minimum sustained-release immediate darkening dose in the area without product application and the minimum sustained-release immediate darkening dose in the area with product application—which was previously judged by two or more skilled examiners—is performed by an artificial intelligence device. As a result, the determination process is automated by artificial intelligence, thereby reducing the time required for determination. At this time, by detecting erythema and darkening by reflecting the subject's skin tone, the determination of UV protection efficacy is objectified, and the reliability of the determination result is increased.

[0086]

[0087] The present invention has been described above through preferred embodiments. However, the aforementioned embodiments are merely illustrative of the technical concept of the present invention, and those skilled in the art will understand that various modifications are possible within the scope of the technical concept of the present invention. Therefore, the scope of protection of the present invention should be interpreted by the matters described in the claims rather than by specific embodiments, and all technical concepts within an equivalent scope should also be interpreted as being included within the scope of rights of the present invention.

[0088]

[0089] Explanation of the symbols

[0090] 100: Image acquisition unit

[0091] 200: Skin tone information output unit

[0092] 300: Symptom detection unit

Claims

1. A method for determining UV blocking efficacy by applying a functional cosmetic to a subject's skin, irradiating it with ultraviolet rays, and observing skin changes to determine blocking efficacy, wherein The process of observing changes in skin irradiated with ultraviolet rays is performed through the detection of erythema or melanosis by an artificial intelligence algorithm on images of the subject's skin, and A method for determining the UV blocking efficacy of an AI-based functional cosmetic, characterized by the above artificial intelligence algorithm performing erythema or darkening detection by reflecting skin tone information derived from an image of a subject's skin that has not been irradiated with ultraviolet rays.

2. In Claim 1, A method for determining the UV blocking efficacy of an AI-based functional cosmetic, characterized in that the above-mentioned skin tone information is obtained by calculating the average RGB value of an image of the skin of a subject not irradiated with ultraviolet rays, converting it into an L*, a*, b* color space, and calculating the Individual Typology Angle (ITA) for the angle formed by the L and b axes.

3. In Claim 1, A method for determining the UV blocking efficacy of an AI-based functional cosmetic, characterized in that the above-mentioned skin tone information is obtained by converting RGB into L*, u*, v* color spaces and calculating a UV histogram from an image of the skin of a subject not irradiated with ultraviolet rays.

4. In Claim 1, A method for determining the UV blocking efficacy of an AI-based functional cosmetic, characterized in that the above skin tone information is the average RGB value of an image taken of the skin of a subject not irradiated with ultraviolet rays.

5. A UV protection efficacy determination system applied in the process of determining blocking efficacy by observing skin changes after applying a functional cosmetic to a subject's skin and irradiating it with ultraviolet rays, wherein An image acquisition unit that acquires an image of the subject's skin; A skin tone information calculation unit that calculates information on skin tone from an image of skin not irradiated with ultraviolet rays among images of a subject's skin; and It includes a symptom detection unit that detects erythema or melanosis in the UV-irradiated area among images of the subject's skin, and A system for determining the UV blocking efficacy of an AI-based functional cosmetic, characterized in that the above symptom detection unit detects erythema or darkening in a UV-irradiated area irradiated with UV rays by an artificial intelligence algorithm, and the above artificial intelligence algorithm detects erythema or darkening by reflecting information on skin tone calculated by the above skin tone information calculation unit.

6. In Claim 5, The above-described skin tone information calculation unit is characterized by calculating the average RGB value of an image of a subject's skin that has not been irradiated with ultraviolet rays, converting it into L*, a*, b* color spaces, and calculating the Individual Typology Angle (ITA) for the angle formed by the L and b axes, in an AI-based system for determining the UV blocking efficacy of functional cosmetics.

7. In Claim 5, The above-described skin tone information calculation unit is characterized by converting RGB into L*, u*, and v* color spaces in an image of a subject's skin that has not been irradiated with ultraviolet rays, and calculating a UV histogram, in an AI-based system for determining the UV blocking efficacy of functional cosmetics.

8. In Claim 5, The above-mentioned skin tone information calculation unit is characterized by calculating the average RGB value of an image of a subject's skin that has not been irradiated with ultraviolet rays, in an AI-based system for determining the UV blocking efficacy of functional cosmetics.

Images