KR102465728B1 - Method and apparatus for detecting dark circles and quantifying the degree of dark circles - Google Patents

Abstract

According to various embodiments, a device for diagnosing dark circles comprises: a memory; and at least one processor operatively connected to the memory. The at least one processor can perform the operations of: being connected to the device to obtain the face image of a user photographed by a camera for photographing the face image of the user; detecting a dark circle area of the user from the obtained face image of the user; calculating a dark circle score for the detected dark circle area; and outputting the calculated dark circle score through a display connected to the device.

Description

Method and apparatus for quantifying the degree of dark circles by detecting dark circles

The present invention relates to a method and apparatus for detecting dark circles and quantifying the degree of dark circles, and more particularly, by using an image of a user obtained by photographing a user to subdivide a user's dark circles using artificial intelligence, It relates to a method and apparatus for quantifying the degree of dark circles.

Unless otherwise indicated herein, material described in this section is not prior art to the claims in this application, and inclusion in this section is not an admission that it is prior art.

Dark circles refer to a disease in which the skin color is darkly discolored due to fat under the eyes, blood vessels exposed by thin skin, and pigmentation around the eyes. The cause is caused by various environments such as lack of sleep, stress, and external environment. Dark circles can quickly change and deteriorate depending on the external and internal environment, and it is necessary to periodically check the appearance and size of your dark circles and take corrective measures.

Accordingly, various compositions, cosmetics, patches, masks, care devices, etc. for improving dark circles are currently being released. However, it is important to accurately diagnose the user's dark circle in order to recommend and present the composition, cosmetics, patch, mask care device, etc. for improving the dark circle as described above to the user. However, currently, only the presence or absence of dark circles is used as a criterion for diagnosing dark circles in order to measure the user's skin condition, and a method for quantifying dark circles is insufficient.

Publication No. 10-2021-0077533 (2021.06.25., published) Patent Publication No. 10-2014-0112046 (2014.09.22., published) Patent Publication No. 10-2022-0078231 (2022.06.10., published)

An object of the present invention to solve the above problems is to provide a method and apparatus for quantifying the degree of dark circles by detecting dark circles.

According to various embodiments of the present disclosure, a device for diagnosing dark circles includes a memory and at least one processor operatively connected to the memory, and the at least one processor works in conjunction with the device to capture a user's face image. obtaining the face image of the user photographed through a camera; detecting a user's dark circle area from the obtained user's face image; calculating a dark circle score for the detected dark circle area; and an operation of outputting the calculated dark circle score through a display interlocked with the device.

Converting, by the at least one processor, pixel values included in the dark circle area into probability values; calculating the dark circle area based on the number of pixels for which the converted probability value exceeds a preset threshold value; and calculating the dark circle score based on the calculated dark circle area.

performing, by the at least one processor, histogram equalization on the face image based on the face image and a preset reference image; and an operation of converting RGB values of pixels included in the face image on which histogram equalization has been performed into a hemoglobin distribution.

The at least one processor may perform an operation of converting the face image of the user into an HSV format image through a color conversion algorithm.

The at least one processor calculates the dark circle score based on the dark circle area, the ratio of the dark circle area to the user's face area, the V value of the face image converted to the HSV format image, and the V-S value. can

According to various embodiments disclosed in this document, not only the position and range of the dark circle are measured, but also the degree of the corresponding area can be measured to accurately diagnose the user's dark circle.

In addition to this, various effects identified directly or indirectly through this document may be provided.

1 is a diagram illustrating a dark circle diagnosis system according to an exemplary embodiment.
2 is a diagram showing the main components of a user terminal.
3 is a diagram related to diagnosing dark circles by performing histogram equalization on a face image.
4 is a diagram related to quantifying a user's dark circle or calculating a dark circle score according to an embodiment.
5 is a diagram showing a hardware configuration of a user terminal according to FIG. 1 .

Since the present invention can make various changes and have various embodiments, specific embodiments will be illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. Like reference numerals have been used for like elements throughout the description of each figure.

Terms such as first, second, A, and B may be used to describe various components, but the components should not be limited by the terms. These terms are only used for the purpose of distinguishing one component from another. For example, a first element may be termed a second element, and similarly, a second element may be termed a first element, without departing from the scope of the present invention. The terms and/or include any combination of a plurality of related recited items or any of a plurality of related recited items.

It is understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle. It should be. On the other hand, when an element is referred to as “directly connected” or “directly connected” to another element, it should be understood that no other element exists in the middle.

Terms used in this application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, the terms "include" or "have" are intended to designate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, but one or more other features It should be understood that the presence or addition of numbers, steps, operations, components, parts, or combinations thereof is not precluded.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in the present application, they should not be interpreted in an ideal or excessively formal meaning. don't

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a dark circle diagnosis system 10 according to an embodiment. Referring to FIG. 1 , the dark circle diagnosis system 10 may include a user terminal 100, a database 200, and the like. Operations described below may be performed or implemented through a platform (eg, a web page and/or application) controlled by the user terminal 100 . In other words, the user terminal 100 may provide a website where the user can access various servers through a network using the user terminal 100 to input, register, and output various information, and the user By being installed and executed in the terminal 100, an application capable of inputting, registering, and outputting various information can be provided.

The dark circle diagnosis system 10 may detect a dark circle area in the user's face image using the user's face image, and perform dark circle quantification on the detected dark circle area. The dark circle diagnosis system 10 trains a dark circle detection model using various people's face image training datasets stored in the database 200, and detects a user's dark circle area based on the learned dark circle detection model. Detection or quantification can be performed for

The user terminal 100 includes a communicable desktop computer, a laptop computer, a notebook, a smart phone, a tablet PC, a mobile phone, and a smart watch. (smart watch), smart glass, e-book reader, PMP (portable multimedia player), portable game console, navigation device, digital camera, DMB (digital multimedia broadcasting) player, digital voice It may be a digital audio recorder, a digital audio player, a digital video recorder, a digital video player, and a personal digital assistant (PDA).

The user terminal 100 may be a server that performs operations described in FIGS. 2 to 4 as one or more servers as well as a portable device. The user terminals 100 are each connected to a communication network and may transmit and receive data to and from each other through the communication network. For example, communication networks include Local Area Network (LAN), Metropolitan Area Network (MAN), Global System for Mobile Network (GSM), Enhanced Data GSM Environment (EDGE), High Speed Downlink Packet Access (HSDPA), and W-CDMA. (Wideband Code Division Multiple Access), CDMA (Code Division Multiple Access), TDMA (Time Division Multiple Access), Bluetooth, Zigbee, Wi-Fi, Voice over Internet Protocol (VoIP) , LTE Advanced, IEEE802.16m, WirelessMAN-Advanced, HSPA+, 3GPP Long Term Evolution (LTE), Mobile WiMAX (IEEE 802.16e), UMB (formerly EV-DO Rev. C), Flash-OFDM, iBurst and MBWA (IEEE Various types of wired or wireless networks can be used, such as 802.20) systems, HIPERMAN, Beam-Division Multiple Access (BDMA), Wi-MAX (World Interoperability for Microwave Access), and 5G.

2 is a diagram showing the main components of the user terminal 100 .

Referring to FIG. 2 , functional modules of a user terminal (eg, processor 110) 100 may include a skin diagnosis unit 101, a dark circle detection unit 102, and a dark circle diagnosis unit 103.

The skin diagnosis unit 101 selects a user's face image captured by a camera (eg, a camera of the user terminal 100 and/or a camera of another skin measuring device) and/or images transmitted through a wired/wireless network. A user's face image including the user's face may be obtained. The photographed face image of the user may include distance data. The user's face image may be an image of a facial area set by a distance sensor (eg, ToF sensor) and/or a landmark automatically set based on AI based on the camera.

The skin diagnosis unit 101 pre-processes the face image by correcting color values (eg, RGB pixel values) of the face image by smoothing color and removing noise through a lighting and color regularization algorithm. can

The skin diagnosis unit 101 may determine a landmark of the user's face image captured based on the distance sensor and/or the camera. The skin diagnosis unit 101 may detect the user's face region based on the landmark, and determine the user's skin characteristics based on the detected face region. Skin characteristics may include skin type and skin condition. For example, skin characteristics may include skin moisture, oil content, sebum, pH level, sensitivity, elasticity, wrinkles, skin color/skin tone, pore condition, pigmentation, and dead skin condition. The skin diagnosis unit 101 may determine the user's skin characteristics based on the facial image obtained through the user terminal 200 and/or other skin measurement devices. That is, the skin diagnosis unit 101 may determine the user's skin characteristics such as moisture, oil, sebum, pH, sensitivity, wrinkles, skin tone, pore condition, pigmentation, and dead skin condition of the user's skin based on the face image. have. A skin type of the user may be determined based on the determined skin characteristics.

The dark circle detection unit 102 may select a dark circle area more limitedly and determine a dark circle expected area in order to detect and diagnose the dark circle within the limited area. The dark circle detection unit 102 detects the left eye and the right eye in the acquired face image, determines the outermost line of each of the detected left eye and right eye, and bends at a predetermined angle or more from the determined outermost line. point can be determined. For example, the dark circle detection unit 102 determines a first point and a second point bent at a preset angle or more in the first outermost line corresponding to the left eye, and determines the second outermost line corresponding to the right eye. It is possible to determine a third point and a fourth point that are bent at an angle greater than or equal to a preset angle.

The dark circle detector 102 may calculate a first distance between the first point and the second point, and calculate a second distance between the third point and the fourth point. The dark circle detection unit 102 includes a first area from a first lowermost point of the first outermost line corresponding to the left eye to a point away from the bottom by a first distance and a second outermost line corresponding to the right eye. Each of the second regions from the lowest point to a point away from the bottom by a second distance may be determined as the expected dark circle region.

The dark circle detection unit 102 may perform supervised learning or unsupervised learning on the dark circle detection model, and the dark circle detection model may be implemented as an artificial neural network (ANN). An artificial neural network is a predictive model implemented in software or hardware that mimics the computational power of a biological system using a large number of artificial neurons (or nodes).

The dark circle detection unit 102 may acquire a plurality of cropped face images based on facial landmarks in order to learn an artificial neural network model (eg, a dark circle detection model). The dark circle detection unit 102 may classify the plurality of face images for learning an artificial neural network model into a training data set and a verification data set at a ratio of 1:9. The training data set can be increased by a factor of 10 through the cutmix technique. An artificial neural network model can extract features from a face image using a Swin (shifted window) Transformer built on the basis of a convolutional neural network. The dark circle detection unit 102 may store the extracted features in a memory. The dark circle detection unit 102 may determine the dark circle area through the extracted features and a Unified Perceptual Parsing Network (UperNet).

The dark circle detection unit 102 may convert pixel values included in the determined dark circle area into probability values through Equation 1 below. The dark circle detection unit 102 may calculate a dark circle area based on the sum of the number of pixels exceeding a preset threshold among the converted probability values, and calculate a dark circle score based on the calculated dark circle area.

The dark circle detection unit 102 may generate a binarized map by decomposing into (0, 1) based on a threshold value of 0.5 in order to infer a specific location of the dark circle.

은 다크서클 영역에 포함된 픽셀 값을 의미할 수 있다. 구체적으로,

은 인공신경망 모델의 출력 중 다크서클 영역에 포함된 픽셀 값을 의미할 수 있다.In Equation 1, S _area is the dark circle area,

may mean a pixel value included in the dark circle area. Specifically,

may mean a pixel value included in the dark circle area among outputs of the artificial neural network model.

The dark circle detection unit 102 may calculate an average value of pixel values included in the dark circle area and a standard deviation of the pixel values based on the detected location of the dark circle area.

The dark circle detection unit 102 may convert the user's face image into an HSV format image through a color conversion algorithm. In other words, since the distribution of HSV values is relatively easier to cluster than values corresponding to the RGB color space, the dark circle detection unit 102 converts the user's RGB format face image into an HSV format image through a color conversion algorithm can do. The dark circle detection unit 102 determines the first dark circle (eg, black dark circle) as the value of V of the pixel value included in the dark circle area is lower, and the higher the S value of the pixel value included in the dark circle area is. It can be determined as a second dark circle (eg, a brown dark circle).

The dark circle detection unit 102 may standardize the dark circle area corresponding to the dark circle area calculated based on the threshold in the face image converted to the HSV format image through Equation 2 below.

(뮤)는 획득된 상기 평균값,

(시그마)는 획득된 상기 표준편차를 의미할 수 있다.In Equation 2, Z is a standardized value (normalized value), X is a value corresponding to each pixel,

(mu) is the average value obtained,

(Sigma) may mean the standard deviation obtained.

The dark circle detection unit 102 may interpolate the standardized value z into values in the range [1, 2]. The interpolated value may be applied as a weight for the S _area calculated by Equation 1 above.

3 is a diagram related to diagnosing dark circles by performing histogram equalization on a face image.

In order to more accurately identify dark circles, the dark circle detection unit 102 may convert the RGB values of each pixel included in the user's face image into a hemoglobin distribution through Equation 3 below. However, since the hemoglobin distribution is very sensitive to the white balance of the captured image, the dark circle detection unit 102 applies histogram equalization to the user's face image as a preprocessing to correct the color of the image as a whole.

Referring to FIG. 3 , since the skin tone of user A and user B are different, when diagnosing dark circles in the same way, results that cannot be compared with the same standard as image A 310 and image B 320 will be derived. Therefore, it is necessary to convert image B 320 into image B' 330 through histogram equalization. The dark circle detection unit 102 may generate a reference image for histogram smoothing by calculating a histogram in the [0, 255] range of the images of the training data set and then inserting a value corresponding to an average to each pixel.

In Equation 3, HD may mean Hemoglobin Distribution, and I may mean a pixel value corresponding to each pixel. For example, in Hist(x, x_ref), the input image may mean x and the reference image may mean x_ref.

In Equation 4, H _h is a value corresponding to hemoglobin, HD (G) is a Hemoglobin Distribution calculated based on the G (green) value corresponding to each pixel, HD (R) is R corresponding to each pixel ( red) may mean a hemoglobin distribution calculated based on the value.

In Equation 5, H _mh is a value corresponding to the melanin pigment, HD (G) is the Hemoglobin Distribution calculated based on the G (green) value corresponding to each pixel, HD (R) is R corresponding to each pixel The Hemoglobin Distribution HD (B) calculated based on the (red) value may mean the Hemoglobin Distribution calculated based on the B (blue) value corresponding to each pixel.

The dark circle detection unit 102 interpolates both H _h and H _mh in the [1, 2] range, and uses an average value of pixels corresponding to the dark circle area as a weight through an artificial neural network.

4 is a diagram related to quantifying a user's dark circle or calculating a dark circle score according to an embodiment.

The dark circle diagnosis unit 103 determines the area of the dark circle, the ratio of the area of the user's dark circle to the size of the user's eyes, the degree of the first dark circle (eg black dark circle), the second dark circle (eg brown dark circle) ), the degree of hemoglobin, and the degree of melanin pigment can be used to calculate the dark circle score. The dark circle diagnosis unit 103 may calculate an area ratio of the user's dark circle to the size of the user's eyes. In other words, even if the same area of the dark circle is detected, the dark circle may be relatively severe or weak depending on the size of the user's eyes.

The dark circle diagnosis unit 103 may calculate the dark circle score through Equation 6 below.

)는 V 표준화값에 대한 가중치, (

)는 V-S 표준화값에 대한 가중치, (

)는 헤모글로빈에 대한 가중치, (

)는 멜라닌 색소에 대한 가중치, k는 다크서클의 면적에 대한 가중치일 수 있다.In Equation 6, S is the dark circle score, S _area is the area of the dark circle, (

) is the weight for the standardized value of V, (

) is the weight for the VS standardized value, (

) is the weight for hemoglobin, (

) may be a weight for the melanin pigment, and k may be a weight for the area of the dark circle.

)은 보간된 [1, 2] 범위의 역수로 계산되어지고, 갈색 다크서클에 대한 가중치 값(

)은 상기 수학식 2에서 계산된 V와 S를 뺀 {V - S}의 값에 보간을 적용한 값을 가중치로 계산될 수 있다.The weight value for the V value corresponding to the deep dark circle (

) is calculated as the reciprocal of the interpolated [1, 2] range, and the weight value for the brown dark circle (

) may be calculated as a weight value obtained by applying interpolation to the value of {V - S} obtained by subtracting V and S calculated in Equation 2 above.

The dark circle score S exists one by one for each input image, which is the user's face image, and the dark circle diagnosis unit 103 uses the Face Descriptor for the same person who repeatedly photographed over several days according to the timeline. A dark circle score may be given by updating the dark circle score.

The dark circle diagnosing unit 103 may calculate a final dark circle score based on the difference in degree of dark circles between the eyes and/or the level of makeup of the user. In other words, if the user's makeup is heavy, there is a high probability that the dark circle diagnosis for the user based on the input image will not be accurate, so the dark circle diagnosis unit 103 can reflect the makeup level to the dark circle score, The circle diagnosis unit 103 may reflect the difference in the degree of the left and right dark circles to the dark circle score considering that the skin condition is inconsistent when the degree difference between the left and right dark circles is severe.

The dark circle diagnosis unit 103 may determine the user's makeup and/or the degree of makeup. The dark circle diagnosis unit 103 may determine the level of makeup of the user based on the user's current face image and the user's bare face image. The current face image may be a user's face image obtained in real time. For example, the current face image may be a face image in which a part of the user's face is detected based on the acquired image. The bare face image may be an image of a face without makeup obtained from a camera linked to a user terminal or a server, and the image of the face without makeup of the user may be an image inferred based on deep learning of an input image including the user's face. can For example, the bare face image may be a face image determined through feature points of the user's face image using an artificial neural network pre-supervised.

The dark circle diagnosis unit 103 compares the user's bare face image with the user's current face image, and based on a change in skin color, a change in the number of troubles (eg, a change in the number of red spots, a change in the number of freckles), etc. You can decide the level of makeup.

The dark circle diagnosis unit 103 may determine, as the bare face image, a facial image having feature points obtained as an output of an artificial neural network among the bare face images of the user stored in the database 200 . Alternatively, the dark circle diagnosis unit 103 may determine, as the average facial image, a facial image most similar to a facial image having feature points obtained as an output of an artificial neural network among facial images of the user stored in the database 200 .

The dark circle diagnosis unit 103 may include training of a bare face image estimation model and a bare face image estimation model. The bare face image estimation model learning unit and the bare face image estimation model are divided according to the functions of the dark circle diagnosis unit 103, and it is obvious that the dark circle diagnosis unit 103 can perform all of the corresponding functions.

The dark circle diagnosis unit 103 may obtain bare face image learning data stored in the database 200 . The bare face image training data according to an embodiment may be training data having a feature point of a bare face image as an input value and an average feature point of the face image as an output value. A feature point of the face image may be a color value corresponding to each pixel.

The bare face image estimation model learning unit may acquire bare face images of users corresponding to each of the plurality of user terminals as training data. The bare face image estimation model learning unit may learn the bare face image estimation model using the acquired training data. An artificial neural network may be used as a bare face image estimation model according to an embodiment. An artificial neural network is a predictive model implemented in software or hardware that mimics the computational power of a biological system using a large number of artificial neurons (or nodes).

The bare face image estimation model may be supervised and learned by the bare face image estimation model learning unit using 'feature points of the bare face image' and 'average feature points of the bare face image'. In this case, supervised learning refers to learning to find an output value according to a given input value by using data having input values and corresponding output values as learning data, and means learning performed in a state where the correct answer is known. The set of inputs and outputs given to supervised learning is called training data. That is, the above-described 'feature points of the bare face image' and 'average feature points of the bare face image' are input values and output values, respectively, and can be used as training data for supervised learning of the bare face image estimation model.

For example, the bare face image estimation model learning unit converts the feature points of the bare face image into a unique first one-hot vector to generate an input value, and converts the average feature point of the bare face image into a unique first one-hot vector. 2 After converting to one-hot vector to generate an output value, the generated input value and output value can be used to supervise the bare face image estimation model. Here, the first one-hot vector and the second one-hot vector may be vectors in which one of the component values constituting the vector is '1' and the other component values are '0'.

In one embodiment, the bare face image estimation model receives an input value, multiplies an input layer having nodes corresponding to the number of components of the first one-hot vector, and a connection strength (or weight) for each output value of the input layer. . , one or more hidden layers for adding and outputting a bias; and an output layer that multiplies each output value of the hidden layer by a connection strength (or weight) and outputs the result using an activation function. For example, the activation function may be a ReLU function or a Softmax function, but is not limited thereto. Connection strength and bias can be continuously updated by supervised learning.

Specifically, the bare face image estimation model may be supervised so that an output value of a loss function according to a given input value (first one-hot vector) and output value (second one-hot vector) is minimized.

The dark circle diagnosis unit 103 may compare the bare face image with the current face image of the user to determine the level of makeup of the user. The dark circle diagnosis unit 103 may determine the level of makeup of the user for each partial area in the user's estimated bare face image and the user's current face image. For example, the dark circle diagnosing unit 103 may determine the user's makeup level for each partial region, such as the user's forehead, nose, eye rims, eyebrows, lips, and cheeks, in the bare face image and the current face image.

In Equation 2, y is an index representing the degree of makeup, p is the number of types of troubles, c is the degree of change in color values, h is the number of partial areas, w1 is a weight for the troubles, and w2 is the skin color A weight for , n may be an index for a bare face image, and m may be an index for a current face image.

The dark circle diagnosis unit 103 may calculate the degree of change in trouble and the degree of color change for each partial region in the bare face image and the current face image, and determine the degree of makeup of the user based on the calculated degree of change in trouble and color change. have. The trouble gradient is calculated based on the difference in the number of detected troubles by comparing the bare face image with the user's current face image, and the color gradient is the R (red pixel) value, G( It can be determined by the green pixel) value and the B (blue pixel) value.

The dark circle diagnosis unit 103 may calculate a final dark circle score based on the difference in degree of dark circle between the left eye and the right eye.

The dark circle diagnosis unit 103 may calculate a first dark circle score corresponding to the left dark circle corresponding to the left eye and a second dark circle score corresponding to the right dark circle corresponding to the right eye. have. The dark circle diagnosis unit 103 may calculate a dark circle score difference between the first dark circle score and the second dark circle score, and calculate a final dark circle score based on the calculated dark circle score difference. The dark circle diagnosis unit 103 may calculate a final dark circle score lower as the difference between dark circle scores increases.

The dark circle diagnosis unit 103 determines, when a higher score of the first dark circle score and the second dark circle score is higher than the preset first score, the dark circle score difference within the first weight range as the dark circle score difference increases. It is possible to determine a higher weight for . The dark circle diagnosis unit 103 determines, when a higher score of the first dark circle score and the second dark circle score is lower than the preset second score, the dark circle score difference within the second weight range as the dark circle score difference increases. It is possible to determine a higher weight for . The lowest weight in the first weight range may be higher than the highest weight in the second weight range. For example, the first weight range may be greater than 0.5 and less than 0.8, and the second weight range may be greater than 0.2 and less than 0.5.

The dark circle diagnosis unit 103 may calculate the final dark circle score based on the dark circle score difference and a weight for the dark circle score difference.

FIG. 5 is a diagram showing a hardware configuration of the user terminal 100 according to FIG. 1 .

Referring to FIG. 5 , the user terminal 100 includes at least one processor 110 and a memory for storing instructions instructing the at least one processor 110 to perform at least one operation. (memory).

The at least one operation may include at least some of the above-described operations or functions of the user terminal 100 and may be implemented in the form of instructions and performed by the processor 110 .

Here, the at least one processor 110 may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor for performing methods according to embodiments of the present invention. can Each of the memory 120 and the storage device 160 may include at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory 120 may be one of a read only memory (ROM) and a random access memory (RAM), and the storage device 160 may be a flash-memory. , a hard disk drive (HDD), a solid state drive (SSD), or various memory cards (eg, a micro SD card).

In addition, the user terminal 100 may include a transceiver 130 that performs communication through a wireless network. In addition, the user terminal 100 may further include an input interface device 140 , an output interface device 150 , a storage device 160 , and the like. Each component included in the user terminal 100 may be connected by a bus 170 to communicate with each other. In FIG. 5, the user terminal 100 has been described as an example, but is not limited thereto. For example, a plurality of user terminals may include a component according to FIG. 5 .

The methods according to the present invention may be implemented in the form of program instructions that can be executed by various computer means and recorded on a computer readable medium. Computer readable media may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on a computer readable medium may be specially designed and configured for the present invention or may be known and usable to those skilled in computer software.

Examples of computer readable media may include hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions may include not only machine language codes generated by a compiler but also high-level language codes that can be executed by a computer using an interpreter and the like. The hardware device described above may be configured to operate with at least one software module to perform the operations of the present invention, and vice versa.

In addition, the above-described method or device may be implemented by combining all or some of its components or functions, or may be implemented separately.

Although the above has been described with reference to preferred embodiments of the present invention, those skilled in the art will variously modify and change the present invention within the scope not departing from the spirit and scope of the present invention described in the claims below. You will understand that it can be done.

100: user terminal 200: database

Claims (5)

In the device for diagnosing dark circles,
a memory and at least one processor operatively coupled with the memory, the at least one processor comprising:
obtaining the face image of the user captured by a camera that captures the face image of the user in conjunction with the device;
detecting a user's dark circle area from the obtained user's face image;
calculating a dark circle score for the detected dark circle area; and
Performing an operation of outputting the calculated dark circle score through a display linked to the device;
The operation of detecting the dark circle area,
converting pixel values included in the dark circle area into probability values;
Calculating a dark circle area based on the number of pixels for which the converted probability value exceeds a preset threshold,
Calculate the dark circle area through the following equation,

In the above equation, S _area is the dark circle area,

Means the probability value corresponding to pixel values included in the dark circle area,
An operation for calculating the dark circle score based on the dark circle area, the ratio of the dark circle area to the user's face area, the V value and the VS value in the face image converted to an HSV format image. delete In claim 1,
The at least one processor,
performing histogram equalization on the face image based on the face image and a preset reference image; and
An apparatus for performing an operation of converting an RGB value of each pixel included in the face image on which histogram equalization has been performed into a hemoglobin distribution. In claim 1,
The at least one processor,
An apparatus for performing an operation of converting the face image of the user into an HSV format image through a color conversion algorithm. delete

Images