KR20200137897A - Method and system for diagnosing skin based on an artificial intelligence - Google Patents

Abstract

Provided are a method and a system for learning data using an artificial neural network and diagnosing skin based thereon. According to the present invention, the skin diagnosis method performed by a computing device comprises the steps of: receiving first measurement data measuring a skin condition of a user; inputting the received first measurement data into an artificial intelligence model to obtain first diagnosis data with regard to a skin state of the user; and quantifying the first diagnosis data to generate a first diagnosis result to provide a user guide for skin care of the user.

Description

Artificial intelligence-based skin diagnosis method and system {METHOD AND SYSTEM FOR DIAGNOSING SKIN BASED ON AN ARTIFICIAL INTELLIGENCE}

The present invention relates to an artificial intelligence-based skin diagnosis method and system, and more particularly, to a method and system for learning data using an artificial neural network and diagnosing skin based thereon.

With the development of information digitalization and data storage technology, a large amount of data has been accumulated, and artificial intelligence technology has been introduced and utilized in various fields. Machine learning, a type of artificial intelligence technology, is a technology that analyzes a large number of input data and classifies objects probabilistically or predicts a value within a specific range. Unlike traditional programming methods that derive results determined by clear rules, machine learning works by empirically analyzing a large number of input data to derive results probabilistically.

Meanwhile, skin diseases such as melasma, acne, atopy, and scalp problems are increasing due to recent environmental and genetic factors, and in conjunction with the development of the beauty industry, the development of various related products and services such as beauty devices and cosmetics devices is active. Is going on. Among them, the field that is receiving attention recently is the field of personalized skin care service. Personalized skin care service is a service that diagnoses individual skin conditions (i.e., pores, wrinkles, pigmentation, oil, trouble, spots, or skin cancer) and recommends cosmetics or skin care devices suitable for them. Whether or not to accurately diagnose is directly connected to the quality of service. Traditionally, methods of consulting a dermatologist or performing a simple test using a beauty device have been used to diagnose an individual's skin condition. However, receiving advice from a dermatologist is costly and expensive, and the method of using a cosmetic device has a problem of low reliability of diagnosis results.

Korean Patent Application Publication No. 10-2019-0033125

The technical problem to be solved by the present invention is to provide a skin diagnosis method and system for quickly and accurately diagnosing an individual's skin condition using artificial intelligence.

The technical problem to be solved by the present invention is to provide a skin diagnosis method and system that minimizes the cost and time required for skin diagnosis using artificial intelligence.

Another technical problem to be solved by the present invention is to provide a skin diagnosis method and system for accurately diagnosing various individual skin types and conditions.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems that are not mentioned will be clearly understood by those skilled in the art from the following description.

In order to solve the above technical problem, a skin diagnosis method performed by a computing device according to an embodiment of the present invention includes receiving first measurement data measuring a user's skin condition, and receiving the received first measurement data. By inputting into an artificial intelligence model, obtaining first diagnostic data for the skin condition of the user and generating a first diagnostic result by quantifying the first diagnostic data so that a user guide for skin management of the user is provided. Includes steps.

In an embodiment, after the user guide is provided, receiving second measurement data obtained by re-measurement of the user's skin condition, inputting the second measurement data into the artificial intelligence model, The method may further include obtaining second diagnostic data for the user, and generating a second diagnostic result quantifying the second diagnostic data so that a new user guide for skin care of the user is provided.

As an embodiment, the new user guide may be generated according to a result of determining suitability of the user guide by comparing the first diagnosis result and the second diagnosis result.

As an embodiment, the artificial intelligence model may be a convolutional neural network model that performs machine learning learning based on a plurality of skin condition data.

In one embodiment, the generating of the first diagnosis result includes extracting skin abnormalities from the first diagnostic data, and different skin abnormalities from the skin abnormalities according to the area or condition of the skin abnormalities. And counting the number of skin abnormalities belonging to the same skin abnormality stage among the skin abnormalities.

As an example, the skin abnormality form may be a pigmentation form.

As an embodiment, the user guide may include a recommended guide for cosmetics, skin care devices, food, or skin care methods for the user.

As an embodiment, the user guide may be obtained by inputting the first diagnosis result into the artificial intelligence model.

In order to solve the above technical problem, a skin diagnosis method performed by a computing device according to another embodiment of the present invention includes measuring a skin condition of a user, and diagnosing the measured skin condition through an artificial intelligence model, Transmitting the measured skin condition as first measurement data and providing a user guide for skin care of the user in which the first diagnosis result of the artificial intelligence model is reflected with respect to the first measurement data, and , The first diagnosis result is data obtained by quantifying the measured skin condition.

In one embodiment, after the user guide is provided, re-measuring the skin condition of the user, transmitting the re-measured skin condition as second measurement data, and the artificial intelligence model for the second measurement data Providing a new user guide for skin care of the user in which the second diagnosis result of is reflected, and the second diagnosis result may be data obtained by quantifying the remeasured skin condition.

As an embodiment, the new user guide may be generated according to a result of determining suitability of the user guide by comparing the first diagnosis result and the second diagnosis result.

In one embodiment, the first diagnosis result extracts skin abnormal forms from the output data of the artificial intelligence model for the first measurement data, and different skin abnormal forms according to the area or state of the skin abnormal forms. It may be generated by dividing into skin abnormalities and counting the number of skin abnormalities that belong to the same skin abnormality among the skin abnormalities.

As an embodiment, the user guide may include a recommended guide for cosmetics, skin care devices, food, or skin care methods for the user.

In order to solve the above technical problem, a computing device according to another embodiment of the present invention includes a memory in which a skin diagnostic program is loaded and a processor that executes a skin diagnostic program loaded in the memory, wherein the skin diagnostic program is Instructions for receiving first measurement data measuring a user's skin condition from a diagnosis device, instructions for inputting the first measurement data into an artificial intelligence model to obtain first diagnosis data for the user's skin condition, and skin care of the user And an instruction for generating a first diagnosis result by quantifying the first diagnosis data so that a user guide for the device is provided.

As an embodiment, the skin diagnosis program includes an instruction for receiving second measurement data obtained by re-measurement of the user's skin condition after the user guide is provided, and inputs the second measurement data to the artificial intelligence model. An instruction for obtaining second diagnostic data for a user's skin condition and an instruction for generating a second diagnostic result quantifying the second diagnostic data so that a new user guide for skin care of the user is provided. .

As an embodiment, the new user guide may be generated according to a result of determining suitability of the user guide by comparing the first diagnosis result and the second diagnosis result.

According to various embodiments of the present invention described above, since accurate skin diagnosis is performed on various skin types and conditions for each individual using an artificial intelligence model, it is possible to effectively provide a personalized skin care service suitable for an individual's skin condition.

In addition, since an accurate personalized skin care guide can be provided without undergoing a skin diagnosis from a skin expert or skin specialist, the time and cost required for personalized skin care can be drastically reduced.

In addition, after providing a skin care service to a user, it can be tracked and diagnosed to verify the validity of the existing skin care service and provide improved skin care service accordingly, thereby providing an improved personalized skin care service. .

The effects of the present invention are not limited to the above-mentioned effects, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a diagram illustrating a skin diagnosis system 1000 according to exemplary embodiments.
FIG. 2 is a diagram illustrating a method for machine learning learning by using a plurality of skin diagnostic devices by the machine learning server 200 according to an embodiment of the present invention.
3 is a flow chart showing a skin diagnosis method according to an embodiment of the present invention.
4 is a flowchart illustrating a method of providing a skin diagnosis service to a user according to an embodiment of the present invention.
5 is a flowchart illustrating a method of providing a skin diagnosis service to a user according to another embodiment of the present invention.
6 is a diagram illustrating a method of quantifying a skin diagnosis result according to an embodiment of the present invention.
7 is a histogram exemplarily showing a skin diagnosis result quantified according to an embodiment of the present invention.
8 is a diagram illustrating an example of information provided as a user guide according to an embodiment of the present invention.
9 is an exemplary diagram showing a result of diagnosing a skin condition of a user according to embodiments of the present invention.
10 is a hardware configuration diagram of a skin diagnosis system according to another embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Advantages and features of the present invention, and a method of achieving them will become apparent with reference to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in a variety of different forms, and only these embodiments make the disclosure of the present invention complete, and are common knowledge in the technical field to which the present invention pertains. It is provided to completely inform the scope of the invention to those who have, and the invention is only defined by the scope of the claims. The same reference numerals refer to the same elements throughout the specification.

Unless otherwise defined, all terms (including technical and scientific terms) used in the present specification may be used with meanings that can be commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not interpreted ideally or excessively unless explicitly defined specifically. The terms used in the present specification are for describing exemplary embodiments and are not intended to limit the present invention. In this specification, the singular form also includes the plural form unless specifically stated in the phrase.

As used herein, "comprises" and/or "comprising" refers to the presence of one or more other components, steps, actions and/or elements in which the recited elements, steps, actions and/or elements Or does not exclude additions.

In addition, in describing the constituent elements of the present invention, terms such as first and second may be used. These terms are only used to distinguish the component from other components, and the nature, order, or order of the component is not limited by the term.

Hereinafter, some embodiments of the present invention will be described with reference to the drawings.

1 is a diagram illustrating a skin diagnosis system according to embodiments of the present invention. Referring to FIG. 1, a skin diagnosis system 1000 includes at least one skin diagnosis apparatus 100 and a machine learning server 200.

The skin diagnosis apparatus 100 is a device that measures a user's condition. The skin diagnosis apparatus 100 may include at least one sensor for scanning a user's skin or detecting moisture or oil. The skin diagnosis apparatus 100 may be designed to be suitable for measuring a specific body part, and may be configured to have different sensors according to the body part to be measured.

Meanwhile, although one skin diagnosis apparatus 100 is illustrated in FIG. 1, this is only an example, and the skin diagnosis system 1000 may include two or more skin diagnosis apparatuses. In this case, each skin diagnosis apparatus may be configured to measure skin conditions of different body parts.

The skin diagnosis apparatus 100 measures the user's skin, and transmits the measured result to the machine learning server 200 as measurement data 10. The measurement data 10 transmitted by the skin diagnosis apparatus 100 may include image data photographing or scanning the user's skin, measurement data indicating oil content and moisture saturation of the user's skin, data recording the reactivity of the user's skin to a specific substance, or Other data indicating skin conditions may be included. As an embodiment, the measurement data 10 is image data photographing or scanning the user's skin, measurement data indicating oil content and moisture saturation of the user's skin, data recording the reactivity of the user's skin to a specific substance, or other skin conditions. Derivative data that has been modified or edited may be included.

The machine learning server 200 trains an artificial intelligence model for skin condition diagnosis, receives the measurement data 10 transmitted from the skin diagnosis apparatus 100, inputs it into the learned artificial intelligence model, and Generates a skin diagnosis result (20) for.

Specifically, the machine learning server 200 first collects various skin condition images as training data. Then, based on the collected training data, labeling data is created for them, and the collected data image and labeling data are provided to an artificial intelligence model to perform machine learning learning. In addition, after performing machine learning training, new data that has not been used for machine learning training is input into the artificial intelligence model to verify that machine learning training has actually been performed well, and the training data and labeling data are reinforced according to the results. Machine learning learning can be advanced in a way that continues learning learning.

In addition, when the machine learning server 200 receives the measurement data 10 from the skin diagnosis apparatus 100, it inputs the measured data 10 into the learned artificial intelligence model to obtain diagnostic data that diagnoses the user's skin condition. Then, the skin diagnosis result 20 is generated based on the obtained diagnosis data, and the skin diagnosis result 20 is transmitted to the skin diagnosis apparatus 100 so that the skin diagnosis apparatus 100 can provide a user guide.

As an embodiment, the machine learning server 200 may quantify diagnostic data output from an artificial intelligence model and transmit it as a skin diagnosis result 20. Alternatively, if the diagnostic data output from the artificial intelligence model is already quantified data, the machine learning server 200 transmits the diagnostic data as it is as the skin diagnosis result 20, or transmits only the format in a format suitable for transmission. I can. Details of the quantified skin diagnosis result 20 will be described later in detail with reference to FIGS. 6 and 7.

As an embodiment, the skin diagnosis apparatus 100 that receives the skin diagnosis result 20 may be configured as separate hardware from the skin diagnosis apparatus 100 that previously transmits the measurement data 10. That is, the skin diagnosis apparatus 100 may include a plurality of hardware devices, and measure the user's skin through some of them and transmit the result as measurement data 10, while a machine learning server through another part. The skin diagnosis result 20 of 200 may be received and a user guide for skin management of the user may be provided.

Meanwhile, various artificial intelligence models used by the machine learning server 200 may be used. For example, the machine learning server 100 may use Convolutional Neural Networks (“CNN”) or Recurrent Neural Network (“RNN”) as its artificial intelligence model. CNN is a deep learning model that combines an artificial neural network with filter technology, and is optimized so that the artificial neural network can better learn the characteristics of the input data. CNN shows particularly excellent performance in the field of analyzing and classifying image data. RNN is a deep learning model designed to be suitable for learning data that changes with the passage of time, such as time series data, and is constructed by connecting a network at the reference point (t) and the next point (t+1). It is an artificial neural network. The cyclic artificial neural network has a characteristic that the connection between nodes has a cyclic structure, and shows particularly excellent performance in processing time-series data having time-varying characteristics such as handwriting recognition and speech recognition.

As an embodiment, the artificial paper used by the machine learning server 200 may be a CNN. One of the easiest ways to measure the user's skin condition is to obtain image data by photographing or scanning the user's skin. Also, the image data obtained in this way can contain various information on the skin condition in a compressed manner. The data 10 is often obtained in the form of image data. Therefore, a CNN, which has an advantage in analyzing image data, can be used as an artificial intelligence model.

In this case, the artificial intelligence model of the machine learning server 200 may include an artificial neural network of a graph structure consisting of a plurality of layers and a plurality of nodes constituting each layer, and the plurality of layers includes at least one input layer and at least one It may include a hidden layer and one or more output layers. The input layer refers to the layer that receives the data to be analyzed/learned in the layer structure of the artificial neural network, and the output layer refers to the layer where the result value is output from the layer structure of the artificial neural network. The hidden layer refers to all layers except the input layer and the output layer in the layer structure of the artificial neural network. An artificial neural network is made up of successive layers of neurons, and each layer of neurons is connected to the next layer of neurons. If the input layer and the output layer are directly connected without a hidden layer, each input independently contributes to the output regardless of other inputs, making it difficult to obtain accurate results. In practice, input data is interdependent and combined with each other to affect the output in a complex structure, so by adding a hidden layer, neurons in the hidden layer capture subtle interactions between the inputs that affect the final output.

As an embodiment, the skin diagnosis system 1000 measures a user's skin condition, derives a skin diagnosis result, and recommends cosmetics suitable for the user's skin condition, a skin care method, or a skin care device based on the user guide After providing, it is possible to re-measure the user's skin condition at regular intervals and re-diagnose through an artificial intelligence model, thereby providing an improved user guide as a personalized skin care service. That is, when the first user guide is provided and the user performs skin care accordingly, if the user's skin condition is measured and diagnosed again over time, it is possible to determine whether the first provided user guide was effective for the user. Thus, it is possible to provide a skin care guide most suitable for the user as a personalized skin care service by determining whether to continue the first user guide or to provide a new user guide supplementing the improvement points according to the determination result. A method of providing a two-step diagnosis and a user guide will be described in more detail later in the description of FIGS. 4 and 5.

According to the embodiments of the present invention described in FIG. 1, since accurate skin diagnosis is performed for various individual skin types and conditions using an artificial intelligence model, it is possible to effectively provide a personalized skin care service suitable for an individual's skin condition. .

In addition, since an accurate personalized skin care guide can be provided without undergoing a skin diagnosis from a skin expert or skin specialist, the time and cost required for personalized skin care can be drastically reduced.

In addition, after providing a skin care service to a user, it can be tracked and diagnosed to verify the validity of the existing skin care service and provide improved skin care service accordingly, thereby providing an improved personalized skin care service. .

FIG. 2 is a diagram illustrating a method for machine learning learning by using a plurality of skin diagnostic devices by the machine learning server 200 according to an embodiment of the present invention. Referring to FIG. 2, the skin diagnosis apparatus 100 may include a plurality of skin diagnosis apparatuses, and the machine learning server 200 receives, stores, and stores a plurality of measurement data from a plurality of skin diagnosis apparatuses. Machine learning learning can be performed.

As an embodiment, a plurality of skin diagnosis apparatuses included in the skin diagnosis apparatus 100 may measure skin conditions of different body parts and transmit the measured results to the machine learning server 200. For example, the first skin diagnosis apparatus measures the condition of the facial skin and transmits the result as measurement data 10 to the machine learning server 200, and the second skin diagnosis apparatus measures the condition of the skin on the back of the hand. Is transmitted to the machine learning server 200 as measurement data 10, and the n-th skin diagnosis apparatus may measure the skin condition of the neck and transmit the result to the machine learning server 200 as measurement data 10.

As an embodiment, the machine learning server 200 may classify and label measurement data 10 received from a plurality of skin diagnosis apparatuses according to corresponding body parts, accumulate them, and use them as training data. Thereby, the machine learning server 200 can perform machine learning learning on various individual body parts, and the learned artificial intelligence model can diagnose skin conditions for various body parts.

Hereinafter, each operation of the skin diagnosis method according to embodiments of the present specification will be described. It is to be noted that the subject of each movement is not mentioned or the movement expressed in the passive voice is the subject of the skin diagnosis system 1000.

3 is a flow chart showing a skin diagnosis method according to an embodiment of the present invention. Referring to Figure 3, the skin diagnosis method according to the present invention includes steps S110 to S160.

In step S110, an artificial intelligence model for skin diagnosis is trained through machine learning. The machine learning server 200 first collects various skin condition images as training data. Then, based on the collected training data, labeling data is created for them, and the collected data image and labeling data are provided to an artificial intelligence model to perform machine learning learning. In addition, after performing machine learning training, new data that has not been used for machine learning training is input into the artificial intelligence model to verify that machine learning training has actually been performed well, and the training data and labeling data are reinforced according to the results. Machine learning learning can be advanced in a way that continues learning learning.

As an embodiment, the artificial intelligence model learned in step S110 may be a CNN model.

In step S120, the skin condition of the user is measured through the skin diagnosis apparatus 100. The skin diagnosis apparatus 100 may include at least one sensor for scanning a user's skin or detecting moisture or oil. The skin diagnosis apparatus 100 may be designed to be suitable for measuring a specific body part, and may be configured to have different sensors according to the body part to be measured.

As an embodiment, there may be two or more skin diagnosis apparatuses 100. For example, the skin diagnosis system 1000 may include a plurality of skin diagnosis devices, and each skin diagnosis device may be configured to measure skin conditions of different body parts.

In step S130, the measurement data 10 measuring the skin condition of the user is transmitted from the skin diagnosis apparatus 100 to the machine learning server 200. As an embodiment, the transmitted measurement data 10 includes image data photographing or scanning the user's skin, measurement data indicating oil content and moisture saturation of the user's skin, data recording the reactivity of the user's skin to a specific substance, or other It may include data indicating a skin condition, and the like.

In step S140, the machine learning server 200 performs skin diagnosis using the skin diagnosis model. When the machine learning server 200 receives the measurement data 20, it inputs it to the artificial intelligence model learned in step S110 to obtain diagnostic data on the user's skin condition therefrom. At this time, the obtained diagnostic data is data derived by analyzing the user's skin condition by a machine learning method, and includes information indicating the number and degree of pigmentation of the user's skin, the number and degree of wrinkles, the content of oil, or the number and size of pores. . The result of analyzing the user's skin condition through machine learning will be described in more detail later in FIG. 6.

In step S150, a quantitative skin diagnosis result is generated based on the diagnosis data obtained in step S140. The quantified skin diagnosis result means that the user's skin condition is quantified by a predetermined number or grade. For example, a method of calculating the number of pigmentation by dividing the pigmented regions of the user's skin into individual regions, and calculating the progression stage of pigmentation by analyzing the area, state, and shape of the pigmented region. As a result, atypical skin conditions can be quantified with predetermined numbers or symbols. Details of the quantified diagnosis result will be described later in detail with reference to FIGS. 6 and 7.

As an example, the diagnostic data obtained in step S140 may be data in a form that has already been quantified. In this case, in step S150, the skin diagnosis result may be generated by using the quantified diagnosis data as it is, or processing only the format of the quantified diagnosis data in a form suitable for transmission.

In step S160, a user guide for skin care is provided to the user based on the skin diagnosis result. The user guide refers to a personalized skin care guide provided to a user based on the result of diagnosis on the user's skin condition. For example, the user guide provides information on cosmetics and skin care devices suitable for the user's skin condition, information on foods that are good for skin care or useful skin care methods, or recommended information on foods or habits to avoid for skin care. Can include. By using the right cosmetics, skin care devices, food, skin care methods, etc. recommended in the user guide, users can perform personalized skin care suitable for their skin condition.

As an embodiment, the user guide may include information obtained by processing the skin diagnosis result 20 into a histogram, a comprehensive evaluation table, or the like so that the user can easily recognize his or her skin condition.

According to the embodiments of the present invention described in FIG. 3, since accurate skin diagnosis is performed for various skin types and conditions for each individual using an artificial intelligence model, it is possible to effectively provide a personalized skin care service suitable for an individual's skin condition. .

In addition, since an accurate personalized skin care guide can be provided without undergoing a skin diagnosis from a skin expert or skin specialist, the time and cost required for personalized skin care can be drastically reduced.

4 is a flowchart illustrating a method of providing a skin diagnosis service to a user according to an embodiment of the present invention. In the embodiment of FIG. 4, after firstly measuring and diagnosing the user's skin and providing a customized skin care guide to the user, the skin of the user is secondly measured and diagnosed over time, and the skin care provided previously An improved method of skin diagnosis that complements the guide is described.

According to the embodiment of FIG. 4, after the skin condition measurement is performed for the user (S210a), the skin diagnosis apparatus 100 transmits the measured result to the machine learning server 200 as first measurement data (S210b). Then, the machine learning server 200 inputs the transmitted first measurement data into an artificial intelligence model for skin diagnosis, and performs a machine learning diagnosis on the user's skin condition (S220). Then, the machine learning server 200 transmits the result of performing the machine learning diagnosis to the skin diagnosis apparatus 100 as a first diagnosis result (S230). The skin diagnosis apparatus 100 derives a first user guide including cosmetics, skin care devices, food or skin care methods suitable for a user according to the transmitted first diagnosis result (S240a), and the derived first user guide Provides to the user (S240a). The user applies the provided first user guide to his or her skin care method (S250).

Then, after a predetermined time has elapsed, the skin condition of the user is measured again (S260a). The skin diagnosis apparatus 100 transmits the remeasured result as second measurement data to the machine learning server 200 (S260b). The machine learning server 200 inputs the transmitted second measurement data to the artificial intelligence model for skin diagnosis, similar to the above, and performs machine learning diagnosis on the user's skin condition again (S270). Then, the machine learning server 200 transmits the result of performing the machine learning diagnosis again as the second measurement data to the skin diagnosis apparatus 100 as the second diagnosis result (S280). The skin diagnosis apparatus 100 compares and analyzes the transmitted second diagnosis result with the first diagnosis result, and whether the second diagnosis result shows an improved skin condition compared to the first diagnosis result, that is, the first user guide provided previously Determine whether it is appropriate or not. If the second diagnosis result is improved than the first diagnosis result, it means that the first user guide provided previously was appropriate. Therefore, the skin diagnosis apparatus 100 derives a second user guide that does not deviate significantly from the first user guide and provides the user with Provides (S290a, S290b). In this case, the second user guide may be the same as the first user guide, or may be configured in a form that partially supplements an insufficient portion of the first user guide. On the other hand, if the second diagnosis result is not significantly different from the first diagnosis result or worsens, it means that the previously provided first user guide is not suitable. Therefore, the skin diagnosis apparatus 100 finds the wrong part of the first user guide and is Improve or derive a new second user guide with a different approach from the first user guide and provide it to the user (S290a and S290b).

5 is a flowchart illustrating a method of providing a skin diagnosis service to a user according to another embodiment of the present invention. In the embodiment of FIG. 5, similarly to FIG. 4, a method of providing a user with a first skin diagnosis service and then follow-up observation over time to provide a second skin diagnosis service is described. However, the embodiment of FIG. 5 is different from FIG. 4 in that a user guide for providing customized skin management information to a user is generated by the machine learning server 200.

According to the embodiment of FIG. 5, after the skin condition measurement is performed on the user (S310a), the skin diagnosis apparatus 100 transmits the measured result to the machine learning server 200 as first measurement data (S310b). Then, the machine learning server 200 inputs the transmitted first measurement data into an artificial intelligence model for skin diagnosis, performs a machine learning diagnosis on the user's skin condition, and derives a first diagnosis result (S320). Then, the machine learning server 200 applies the artificial intelligence model again to the first diagnosis result to derive a first user guide including cosmetics, skin care devices, food or skin care methods suitable for the user (S330). . Then, the derived first user guide is transmitted to the skin diagnosis apparatus 100 (S340a). Further, the skin diagnosis apparatus 100 bypasses the received first user guide to the user (340b). Then, the user manages his or her own skin in a method recommended by the first user guide (S350). As such, the reason that the machine learning server 200 derives the user guide in FIG. 5 is that machine learning learning can be used to find out which cosmetics or skin care method is most suitable for a specific skin diagnosis result. That is, the machine learning server 200 may use the artificial intelligence model not only when diagnosing a user's skin condition, but also to find a customized skin care method according to the diagnosis result.

Then, after a predetermined time has elapsed, the skin condition of the user is re-measured (S360a). The skin diagnosis apparatus 100 transmits the remeasured result as second measurement data to the machine learning server 200 (S360b). The machine learning server 200 inputs the transmitted second measurement data to the artificial intelligence model for skin diagnosis similar to the above, performs machine learning diagnosis on the user's skin condition again, and derives a second diagnosis result (S370). In addition, the machine learning server 200 applies the artificial intelligence model again to the second diagnosis result to derive a second user guide including cosmetics, skin care devices, food or skin care methods suitable for the user (S380). . At this time, the machine learning server 200 compares and analyzes the first diagnosis result and the second diagnosis result, and whether the second diagnosis result shows an improved skin condition compared to the first diagnosis result, that is, the first user guide provided above is appropriate. Whether it was done or not, judge its suitability. If the second diagnosis result is improved than the first diagnosis result, it means that the first user guide provided previously was appropriate. Therefore, the machine learning server 200 uses an artificial intelligence model to provide a second user who does not deviate significantly from the first user guide. The guide is derived (S380). On the other hand, if the second diagnosis result is not significantly different from the first diagnosis result or worsens, it means that the first user guide provided above is not suitable. Therefore, the machine learning server 200 uses the artificial intelligence model to A second user guide is derived in the direction of finding a wrong part and improving it or finding a new approach (S380). Then, the derived second user guide is transmitted to the skin diagnosis apparatus 100 (S390a). Further, the skin diagnosis apparatus 100 bypasses the received second user guide to the user (340b).

According to the embodiment of FIGS. 4 and 5, after providing a skin care service to the user first, it is followed by a diagnosis to verify the effectiveness of the existing primary skin care service, etc., and provide an improved secondary skin care service accordingly. How to do it is provided. Through this, it is possible to more accurately determine which skin care method is suitable for the user, and accordingly, an improved personalized skin care service can be provided.

Meanwhile, descriptions overlapping with those previously described in FIGS. 1 to 3 in relation to the skin diagnosis apparatus 100 and the rushing machine server 200 are omitted here. Accordingly, it should be understood that parts not described in FIGS. 4 and 5 with respect to the skin diagnosis apparatus 100 and the machine learning server 200 are the same as those described in FIGS. 1 to 3 above.

6 is a diagram illustrating a method of quantifying a skin diagnosis result according to an embodiment of the present invention. In FIG. 6, as an example of the quantification method, a method of quantifying the degree of skin abnormality (here, pigmentation) of the skin is shown.

Referring to FIG. 6, first, pigmentation patterns 42 shown at the bottom of FIG. 6 are extracted from the user's skin diagnosis data. In addition, the pigmented forms 42 are divided into different pigmentation stages 41 according to the pigmented areas or states of the pigmented forms 42 extracted using a machine learning model. In this way, if the pigmentation step 41 is divided for each pigmentation type 42, the degree of pigmentation of the user can be quantified by counting 43 the number of pigmented portions belonging to the same pigmentation step 41. .

7 is a histogram showing a quantified skin diagnosis result according to an embodiment of the present invention. Referring to FIG. 7, the degree of pigmentation quantified in FIG. 6 is shown as a histogram.

In FIG. 7, the horizontal axis represents the pigmentation step, and the vertical axis represents the number of pigmented portions. Through the histogram of FIG. 7, it is possible to intuitively grasp a user's pigmentation state at a glance. For example, along with the detailed information that'the number of pigmented parts with the degree of pigmentation of the second stage is 23','the user has relatively minor pigmentation (stages 1 to 3), and significant pigmentation (8). Steps to 10 steps) can be intuitively grasped the general information that'there are few'.

According to the embodiments of FIGS. 6 and 7, by quantifying and expressing the skin condition diagnosis result, it is possible to express a skin condition that was difficult to clearly express or quantify with specific and clear numerical values. Through this, the user can more objectively and intuitively recognize his or her skin condition.

8 is a diagram illustrating an example of information provided as a user guide according to an embodiment of the present invention. FIG. 8 illustrates examples of more comprehensive and user-friendly results of the previously quantified skin diagnosis. For example, FIG. 8 may be a screen displayed on a terminal to be shown to a user.

Referring to FIG. 8, methods 61 and 64 more comprehensively showing previously quantified data are illustrated. The diagram 61 on the left is a diagram that divides the user's skin condition into four items: pigmentation, wrinkles, oil, and pores, and comprehensively evaluates the user's skin condition for each item. As an embodiment, the diagram 61 may include three areas 62a, 62b, 62c. For example, the first area 62a is a maintenance area, which means that it is sufficient to maintain the user's skin condition as it is, and the second area 62b is a continuous management area where the user's skin condition is good. Although it is not very serious, it means that continuous management is necessary, and the third area 62c is an intensive management area, which means that the user's skin condition is serious and intensive management to improve it is necessary.

For example, it is assumed that the user's skin condition evaluation chart 61 is derived like the condition graph 63. In this case, by looking at where the state graph 63 and the axis of each item meet in the three areas 62a, 62b, 62c, comprehensive information on the user's skin condition can be easily grasped. For example, when the state graph 63 meets the pigmentation axis in the second region 62b (A) and the wrinkle axis in the third region 62c (B), the user's skin condition is pigmented. At a glance, you can see at a glance that it is in a state that continuous management is necessary for the wrinkles and intensive management is necessary for wrinkles. In this way, the skin diagnosis system 1000 may visually and intuitively provide a skin condition diagnosis result and a guide for skin care to the user.

Meanwhile, the diagram 64 on the right shows another example of intuitively displaying the user's skin condition. In the diagram 64, the user's skin condition is divided into four categories: pigmentation, wrinkles, oil and pores. However, in displaying the state, the numerical value 65b scored between 1 point and 100 points and its scale graph 65a are used. For example, looking at the chart (64), it was found that the user's skin condition has a relatively good score of 61 points for pigmentation and pores, but a relatively poor score of 27 points and 37 points for wrinkles and oils. It is visually and intuitively displayed as a scale graph.

As an embodiment, the skin diagnosis system 1000 and the skin diagnosis method of the present invention may be set to recommend appropriate cosmetics, skin care devices, food, and skin care methods to a user based on the numerical score in FIG. 8.

According to the embodiments of FIG. 8, the skin diagnosis system 1000 and the skin diagnosis method of the present invention can provide a user with a diagnosis result for a user's skin condition and a skin care guide in the future very intuitively and simply.

9 is an exemplary diagram showing a result of diagnosing a user's skin condition according to embodiments of the present invention.

Referring to FIG. 9, an image (a) photographing a user's skin using the skin diagnosis apparatus 100 and an image (b) that detects a pigmentation position by analyzing the photographed image based on an artificial intelligence model are shown. Is shown. Then, a superimposed image (c) displayed by merging the captured image (a) and the detected image (b) is shown.

Looking at the superimposed image (c) of FIG. 9, it can be seen that the pigmented portion detected by the naked eye in the captured image (a) and the image (b) detected based on the artificial intelligence model are very accurately matched. By explaining the results of skin condition diagnosis to the user and presenting these results together when providing personalized skin care services, the user can accurately recognize their skin condition and effectively persuade the need for personalized skin care. I can.

10 is a diagram illustrating a hardware configuration of an exemplary computing device 2000 capable of implementing a skin diagnosis method using artificial intelligence according to embodiments of the present invention.

As shown in FIG. 10, an exemplary computing device implementing a skin diagnosis method using artificial intelligence according to embodiments of the present invention includes at least one processor 2100 and a computer program 2310 executed by the processor 2100. A memory 2300 for loading ), a storage 2900 for storing a computer program 2910 loaded in the memory 2300, a communication interface 2700, and a bus 2500 may be included. In FIG. 10, only components related to an embodiment of the present invention are shown. Accordingly, those of ordinary skill in the art to which the present invention belongs may recognize that other general-purpose components may be further included in addition to the components illustrated in FIG. 10.

The processor 2100 controls the overall operation of each component of a computing device capable of implementing a skin diagnosis method using artificial intelligence according to embodiments of the present invention. The processor 2100 includes a CPU (Central Processing Unit), MPU (Micro Processor Unit), MCU (Micro Controller Unit), GPU (Graphic Processing Unit), or any type of processor well known in the technical field of the present invention. Can be. In addition, the processor 2100 may perform an operation on at least one application or program for executing the method according to the embodiments of the present invention.

The storage 2900 may non-temporarily store one or more computer programs. The storage 2900 is a nonvolatile memory such as a ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), flash memory, etc., a hard disk, a removable disk, or in the technical field to which the present invention belongs. It may be configured to include any known computer-readable recording medium. In the storage 2900, a skin diagnosis program 2910 for performing a skin diagnosis method using artificial intelligence according to embodiments of the present invention is stored.

The memory 2300 stores various types of data, commands, and/or information. The memory 2300 may load the skin diagnosis program 2310 from the storage 2900 in order to execute a skin diagnosis method using artificial intelligence according to embodiments of the present invention.

The bus 2500 provides a communication function between components of the computing device 2000. The bus 2500 may be implemented as various types of buses such as an address bus, a data bus, and a control bus.

The communication interface 2700 supports wired/wireless communication of the computing device 2000. In addition, the network interface 2700 may support various communication methods other than Internet communication. To this end, the network interface 270 may be configured to include a communication module well known in the art.

The skin diagnosis programs 2310 and 2910 using artificial intelligence according to the embodiments of the present invention are loaded into the memory 2300, and instructions for causing the processor 2100 to perform the skin diagnosis method according to the embodiments of the present invention. Includes.

Since each instruction of the skin diagnosis programs 2310 and 2910 according to an embodiment of the present invention performs a skin diagnosis method using artificial intelligence described above, a detailed description thereof will be omitted here to avoid repetition of the description.

Meanwhile, the concept of the present invention described with reference to the drawings may be implemented as computer-readable codes on a computer-readable medium. The computer-readable recording medium may be a removable recording medium (CD, DVD, Blu-ray disk, USB storage device, removable hard disk) or a fixed recording medium (ROM, RAM, computer-equipped hard disk). A computer program recorded in a computer-readable recording medium may be transmitted to another computing device through a network such as the Internet and installed in another computing device, thereby being used in another computing device.

Although the operations are illustrated in a specific order in the drawings, it should not be understood that the operations must be executed in the specific order shown or in a sequential order, or all illustrated operations must be executed to obtain a desired result. In certain situations, multitasking and parallel processing may be advantageous. Moreover, the separation of the various components in the embodiments described above should not be understood as requiring such separation, and the program components and systems described are generally integrated together into a single software product or may be packaged into multiple software products. It should be understood that there is.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those of ordinary skill in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features. I can understand. Therefore, it is to be understood that the embodiments described above are illustrative in all respects and not limiting.

Claims (16)

In the skin diagnosis method performed by a computing device,
Receiving first measurement data measuring a user's skin condition;
Inputting the received first measurement data into an artificial intelligence model to obtain first diagnostic data on the skin condition of the user; And
And generating a first diagnosis result by quantifying the first diagnosis data so that a user guide for skin management of the user is provided. The method of claim 1,
Receiving second measurement data obtained by re-measurement of the user's skin condition after the user guide is provided;
Inputting the second measurement data to the artificial intelligence model to obtain second diagnostic data for the skin condition of the user; And
Generating a second diagnosis result obtained by quantifying the second diagnosis data so that a new user guide for skin care of the user is provided. The method of claim 2,
The new user guide,
The skin diagnosis method, which is generated according to a result of determining suitability of the user guide by comparing the first diagnosis result and the second diagnosis result. The method of claim 1,
The artificial intelligence model,
Skin diagnosis method, which is a convolutional neural network model that performs machine learning learning based on a plurality of skin condition data. The method of claim 1,
Generating the first diagnosis result,
Extracting skin abnormal types from the first diagnostic data;
Dividing the skin abnormal forms into different skin abnormal stages according to the area or condition of the skin abnormal forms; And
And counting the number of skin abnormalities that belong to the same skin abnormality stage among the skin abnormalities. The method of claim 5,
The skin abnormal form is a pigmentation form, skin diagnosis method. The method of claim 1,
The user guide,
Skin diagnosis method comprising a recommended guide for cosmetics, skin care devices, food, or skin care methods for the user. The method of claim 1,
The user guide,
Skin diagnosis method obtained by inputting the first diagnosis result into the artificial intelligence model. In the skin diagnosis method performed by a computing device,
Measuring a user's skin condition;
Transmitting the measured skin condition as first measurement data in order to diagnose the measured skin condition through an artificial intelligence model; And
Including the step of providing a user guide for skin care of the user reflecting the first diagnosis result of the artificial intelligence model with respect to the first measurement data,
The first diagnosis result is data obtained by quantifying the measured skin condition. The method of claim 9,
Re-measurement of the user's skin condition after the user guide is provided;
Transmitting the re-measured skin condition as second measurement data; And
Further comprising the step of providing a new user guide for skin care of the user in which the second diagnosis result of the artificial intelligence model is reflected with respect to the second measurement data,
The second diagnosis result is data obtained by quantifying the re-measured skin condition. The method of claim 10,
The new user guide,
The skin diagnosis method, which is generated according to a result of determining suitability of the user guide by comparing the first diagnosis result and the second diagnosis result. The method of claim 9,
The first diagnosis result,
The skin abnormality types are extracted from the output data of the artificial intelligence model for the first measurement data, and the skin abnormality types are divided into different skin abnormality stages according to the area or condition of the skin abnormality types, and the skin abnormality Skin diagnosis method that is generated by counting the number of skin abnormal forms belonging to the same skin abnormal stage among the forms. The method of claim 9,
The user guide,
Skin diagnosis method comprising a recommended guide for cosmetics, skin care devices, food, or skin care methods for the user. A memory in which a skin diagnostic program is loaded; And
Including a processor for executing the skin diagnostic program loaded in the memory,
The skin diagnosis program,
An instruction for receiving first measurement data obtained by measuring a user's skin condition from the skin diagnosis apparatus;
An instruction for inputting the first measurement data into an artificial intelligence model to obtain first diagnostic data for the skin condition of the user; And
The computing device comprising an instruction for generating a first diagnosis result by quantifying the first diagnosis data so that a user guide for skin care of the user is provided. The method of claim 14,
The skin diagnosis program,
An instruction for receiving second measurement data obtained by re-measurement of the user's skin condition after the user guide is provided;
An instruction for inputting the second measurement data to the artificial intelligence model to obtain second diagnostic data for the skin condition of the user; And
The computing device further comprising an instruction for generating a second diagnosis result obtained by quantifying the second diagnosis data so that a new user guide for skin care of the user is provided. The method of claim 15,
The new user guide,
The computing device, which is generated according to a result of determining suitability of the user guide by comparing the first diagnosis result and the second diagnosis result.

Images