Electronic apparatus for predicting systemic biological age and operation method thereof

Abstract

An electronic apparatus for predicting systemic biological age, according to an embodiment of the present invention, comprises a memory and a processor that predicts systemic biological age, wherein the processor can acquire an eye image of an examinee, process the eye image by using an artificial neural network model that has been established, and generate a predicted result for the systemic biological age of the examinee.

Description

Electronic device for predicting whole-body biological age and method of operation thereof

[0001] The present invention relates to an electronic device for predicting systemic biological age and a method of operation thereof. More specifically, it relates to an electronic device and a method of operation thereof that utilizes an artificial neural network model to evaluate the condition of multiple organs related to a subject's systemic disease based on eye images, and predicts the subject's systemic biological age based thereon.

[0002] Biological age is an indicator that assesses the level of aging associated with an individual's physical health status, and it is useful for determining whether health is better or worse compared to chronological age. Biological age can be utilized to predict the risk of aging-related diseases or to design personalized health management. Existing methods for assessing biological age have evolved through various approaches.

[0003] Existing biological age assessment methods indirectly estimated biological age by utilizing specific physiological indicators of the body. For example, biological age was calculated by synthesizing various physiological indicators such as blood pressure, body mass index (BMI), cholesterol levels, lung capacity, flexibility, blood test-based biomarkers, and genes. While this method has the advantage of being relatively simple to calculate, it has limitations: it fails to reflect the condition of individual organs, requires invasive tests to derive various physiological indicators, and makes it difficult to comprehensively evaluate the overall physical condition. Furthermore, some tests (such as lung capacity, flexibility, and blood pressure) exhibit significant variability depending on the subject's health status or condition, while others (such as blood pressure, cholesterol, and BMI) show high variability due to medication use or physical changes. Additionally, other tests (such as genetic testing) have the disadvantage of not sensitively reflecting changes in physical health.

[0004] Ocular imaging includes anterior segment imaging and posterior segment imaging. Anterior segment imaging is used to assess the health of anterior structures by photographing the front part of the eye, such as the cornea, iris, lens, and anterior chamber. Posterior segment imaging is used to assess the condition of the back part of the eye by photographing the retina, optic nerve, and blood vessels. Posterior segment imaging is also referred to as fundus imaging or retinal imaging. Fundus imaging plays a crucial role as a biomarker capable of providing various health information to a subject through a non-invasive and convenient method. By accurately capturing and analyzing the condition of the retina, optic nerve, and blood vessels, fundus imaging can provide information related not only to ophthalmic diseases but also to systemic diseases such as diabetes, hypertension, cardiovascular disease, and kidney disease. Above all, it has the advantage of visualizing the blood vessels themselves, capturing blood information at a specific point in time, while simultaneously reflecting changes in health status accumulated in vascular or nervous tissues.

[0005] There is a need to develop technology that overcomes conventional limitations, utilizes ocular imaging as a biomarker to comprehensively analyze not only the condition of individual organs (e.g., heart, kidney, brain, lungs, liver, etc.) but also the overall state (e.g., obesity, diabetes, hypertension, sarcopenia, dyspnea, cognitive impairment, etc.), and calculates a subject's biological age through a non-invasive test that places minimal burden on the subject.

[0006] The present invention aims to provide a technology that utilizes ocular images to comprehensively analyze multiple organs and the state of the entire body, and based on this, predicts systemic biological age.

[0007] In addition, the present invention aims to provide an electronic device capable of predicting the biological age, health status, risk level, etc. of individual organs by comprehensively analyzing multiple organs based on eye images, and a method of operating the same.

[0008] In addition, the present invention aims to provide an electronic device capable of predicting an individual's whole-body biological age by comprehensively analyzing an individual's organ and whole-body condition based on eye images, and a method of operation thereof.

[0009] An electronic device for predicting biological age according to one embodiment of the present invention comprises a memory and a processor for predicting biological age, wherein the processor acquires an eye image of a subject, processes the eye image using a pre-established artificial neural network model, and can generate a prediction result for the subject's whole-body biological age.

[0010] Additionally, the artificial neural network model may include a first model that analyzes inputs to the artificial neural network model to extract key features related to multiple organ states and predicts the risk level for each organ state; and a second model that predicts the individual biological age of each organ based on the risk level for each organ generated from the first model, and determines the whole-body biological age of the subject by integrating the results of the age prediction for each organ.

[0011] In addition, the first model may include a backbone module that extracts key features from the eye image and an individual task module that extracts the risk level of at least one organ from the features extracted from the backbone module.

[0012] In addition, the first model may be characterized by learning feature information related to retinal blood vessel patterns, blood vessel color, blood vessel or tissue damage, tissue thickness, nerve fiber layer damage, and macular damage in the eye image through the backbone module, and converting the feature information into a multidimensional feature vector.

[0013] In addition, the individual task module may be characterized by being configured as a multi-task structure that analyzes each of the multiple organs.

[0014] In addition, the first model may further include individual task modules for extracting the metabolic state of the subject from the feature vector.

[0015] Additionally, the second model may include an organ age prediction module that predicts the biological age of an organ using the risk of at least one organ as an independent variable, and an integrated age prediction module that calculates the whole-body biological age of the subject using a weighting algorithm based on the biological age of the organ.

[0016] In addition, the above organ condition may be characterized by including at least two of the heart, kidney, brain, lung, and liver conditions.

[0017] In addition, the above model input may include additional personal data, and the personal data may be characterized by including at least one of anterior segment images, health data, or demographic data.

[0018] In addition, the artificial neural network model may be characterized by being trained by adding at least one of anterior segment images, other measurement data, or demographic data to the input image.

[0019] A system for predicting an individual's whole-body biological age based on a fundus image according to an embodiment of the present invention may include an age prediction model that receives one or more fundus images as model inputs and processes the fundus images to predict an individual's whole-body biological age, wherein the age prediction model processes the model inputs including one or more fundus images to generate a model output that characterizes the individual's whole-body biological age, and wherein the whole-body biological age is determined based on the individual's organ condition and systemic condition.

[0020] Additionally, the age prediction model may be characterized by including: a first model that analyzes the model input using a neural network model to extract key features related to organ and body conditions and predicts the risk level for said organ and body conditions; and a second model that predicts the individual age for each organ and body condition based on the risk level of each organ and body condition generated from the first model, and determines the individual's body biological age by integrating the age prediction results for each condition.

[0021] Additionally, the above organ condition may include one or more conditions related to the heart, kidneys, brain, lungs, and liver, and the above general condition may include one or more conditions related to obesity, diabetes, dyslipidemia, sarcopenia, dyspnea, cognitive impairment, and hypertension.

[0022] In addition, the above model input may additionally include personal data, and the personal data may be characterized by including at least one of anterior segment images, other measurement data, or demographic data.

[0023] In addition, the neural network may be characterized by being trained using at least one of an input image, an anterior segment image, other measurement data, or demographic data.

[0024] A method of operation of an electronic device according to an embodiment of the present invention comprises: a step of acquiring an eye image of a subject; a step of processing the eye image using a previously constructed artificial neural network model; and a step of generating a prediction result for the subject's whole-body biological age, wherein the whole-body biological age may be characterized by being determined based on the condition of a plurality of organs of an individual related to a systemic disease.

[0025] Additionally, the step of processing the eye image may include: a step of analyzing the input to the artificial neural network model to extract key features related to the conditions of multiple organs; a step of predicting the risk level for the condition of each organ; a step of predicting the individual biological age of each organ based on the risk level; and a step of integrating the results of the biological age prediction for each organ to calculate the whole-body biological age of the subject.

[0026] In addition, the step of extracting the above key features may be characterized by learning feature information related to retinal blood vessel patterns, tissue thickness, and nerve fiber layer damage in the above ocular image, and converting the feature information into a multidimensional feature vector.

[0027] In addition, the method may further include a step of extracting the metabolic state of the subject from the above feature vector.

[0028] In addition, the step of calculating the whole-body biological age of the subject may calculate the whole-body biological age of the subject by utilizing a weighting algorithm based on the biological age of the organ.

[0029] According to one embodiment of the present invention, the condition of a subject's multiple major organs can be predicted based on an eye image, and the subject's whole-body biological age can be calculated based on the predicted multiple organ conditions.

[0030] According to one embodiment of the present invention, the systemic condition affecting the entire body, as well as the health status of major organs, can be simultaneously analyzed based on ocular images to calculate the subject's systemic biological age.

[0031] The present invention utilizes the non-invasive nature of ocular imaging to comprehensively evaluate the organ and overall condition of a subject, and based on this, can calculate a more precise biological age of the whole body. Based on this, it can provide the subject with a comprehensive health assessment that includes their overall metabolic and circulatory functions. As a result, it is possible to establish a precise, customized health management plan based on the health status of each organ, and to provide the subject with specific guidelines for early disease prediction and preventive medical measures, thereby contributing to the improvement of the quality of medical services.

[0032] A brief description of each drawing is provided to help to better understand the drawings cited in the detailed description of the invention.

[0033] FIG. 1 schematically illustrates a system including an electronic device for predicting biological age according to one embodiment of the present invention.

[0034] FIG. 2 is a block diagram showing the configuration of an electronic device for predicting biological age according to one embodiment of the present invention.

[0035] FIG. 3 is a conceptual diagram showing the operation of an electronic device for predicting biological age according to one embodiment of the present invention.

[0036] FIG. 4 is a flowchart illustrating the operation of an electronic device for predicting biological age according to one embodiment of the present invention.

[0037] FIGS. 5a and 5b are conceptual diagrams illustrating the operation of an electronic device for predicting biological age according to an embodiment of the present invention.

[0038] FIGS. 6a and 6b are conceptual diagrams illustrating the operation of an electronic device for predicting biological age according to one embodiment of the present invention.

[0039] FIG. 6c is a conceptual diagram showing an example of analysis information regarding the whole-body biological age of a subject according to one embodiment of the present invention.

[0040] FIGS. 7a and 7b are conceptual diagrams illustrating the operation of an electronic device for predicting biological age according to an embodiment of the present invention.

[0041] FIG. 7c is a conceptual diagram showing an example of analysis information regarding the whole-body biological age of a subject according to one embodiment of the present invention.

[0042] FIGS. 8a and 8b are conceptual diagrams illustrating the operation of an electronic device for predicting biological age according to an embodiment of the present invention.

[0043] FIG. 9 is a conceptual diagram showing the operation of an electronic device for predicting biological age according to one embodiment of the present invention.

[0044] Hereinafter, various embodiments of the present invention are described with reference to the accompanying drawings. The present invention is not limited to specific embodiments and should be understood to include various modifications, equivalents, and / or alternatives of the embodiments of the present invention. In connection with the description of the drawings, similar reference numerals may be used for similar components.

[0045] In this document, expressions such as "have," "can have," "include," or "can include" refer to the existence of the relevant feature (e.g., numerical values, functions, actions, or components, etc.) and do not exclude the existence of additional features.

[0046] In this document, expressions such as “A or B,” “at least one of A or / and B,” or “one or more of A or / and B” may include all possible combinations of items listed together. For example, “A or B,” “at least one of A and B,” or “at least one of A or B” may refer to cases including (1) at least one A, (2) at least one B, or (3) both at least one A and at least one B.

[0047] Expressions such as "first," "second," "first," or "second" used in this document may modify various components regardless of order and / or importance, and are used merely to distinguish one component from another without limiting such components. For example, without departing from the scope of rights set forth in this document, the first component may be named the second component, and similarly, the second component may be renamed the first component.

[0048] As used in this document, the expression "configured to" may be replaced, depending on the context, with, for example, "suitable for," "having the capacity to," "designed to," "adapted to," "made to," or "capable of." The term "configured to" does not necessarily mean "specifically designed to."

[0049] In this document, terms transmitted or received between the terminal(s) and the second electronic device(s), such as “command,” “instruction,” “control information,” “message,” “information,” “data,” “packet,” “data packet,” “intent,” and / or “signal,” may include or refer to humanly perceptible ideas or specific electrical representations (e.g., digital codes / analog physical quantities) without being limited by their expression. It will be obvious to a person skilled in the art to which the invention disclosed in this document pertains that the exemplary expressions listed above may be interpreted in various ways depending on the context in which they are used. In this document, “a is greater than B” means not only simply “a is greater than B” but also includes the meaning “a is equal to or greater than B.”

[0050] The terms used in this document are used merely to describe specific embodiments and are not intended to limit the scope of other embodiments. Singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as generally understood by those skilled in the art described in this document. Terms used in this document that are defined in general dictionaries may be interpreted as having the same or similar meaning as they have in the context of the relevant technology, and are not to be interpreted in an ideal or overly formal sense unless explicitly defined in this document. In some cases, even terms defined in this document may not be interpreted to exclude the embodiments of this document.

[0051] FIG. 1 schematically illustrates a system including an electronic device for predicting biological age according to one embodiment of the present invention.

[0052] Referring to FIG. 1, an electronic device (150, hereinafter referred to as the electronic device) for predicting biological age according to one embodiment of the present invention can communicate with a terminal (110) through a network. For example, the terminal (110) may be a terminal used by a subject. As another example, the terminal (110) may be a medical device (e.g., a fundus camera, etc.) used by a medical professional to examine the subject's body.

[0053] For reference, the terminal (110) may be implemented as a computer capable of connecting to a remote server or terminal via a network. For example, the computer may include a notebook, desktop, laptop, etc. equipped with a web browser. Additionally, the terminal (110) may be implemented as a terminal capable of connecting to a remote server or terminal via a network. For example, the terminal (110) is a wireless communication device that ensures portability and mobility, and may include all kinds of handheld-based wireless communication devices such as navigation, PCS (Personal Communication System), GSM (Global System for Mobile communication), PDC (Personal Digital Cellular), PHS (Personal Handphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet) terminal, smartphone, smartpad, tablet PC, etc.

[0054] The electronic device (150) can be coupled to communicate with at least one terminal (110).

[0055] An electronic device (150) according to one embodiment of the present invention may be implemented as a computer device or a plurality of computer devices that provide commands, code files, content, services, etc. The electronic device (150) may be implemented in the form of a self-developed web page or application (APP) created and operated by an individual, company, or hospital, and in this case, data exchange between the electronic device (150) and the terminal (110) may be performed by a user accessing the web page or application through the terminal (110).

[0056] The electronic device (150) may receive an eye image of the subject from the terminal (110). According to an embodiment of the present invention, the eye image may include an image of the anterior part and an image of the posterior part. Preferably, the eye image may be a posterior eye image including a fundus image and a retinal image. The image of the anterior part may include information regarding the subject's cornea, conjunctiva, sclera, lens, iris, anterior chamber, ciliary body, eyelids, etc. The image of the posterior part may include information regarding major components of the fundus, such as the vitreous humor, retina, macula, optic nerve, choroid, optic nerve head, and retinal blood vessels. According to various embodiments of the present invention, the ocular image may include images captured by optical coherence tomography (OCT), fluorescein angiography (FA), indocyanine green angiography (ICGA), ultrasound, confocal microscopy, hyperspectral imaging of the retina, etc.

[0057] The electronic device (150) can extract the condition of the subject's organs (e.g., risk level) and / or metabolic state based on the eye image provided by the terminal (110). For example, the electronic device (150) can extract individual conditions (risk levels) for various organs such as the heart, brain, kidneys, liver, lungs, and blood vessels based on the eye image. Additionally, the electronic device (150) can extract metabolic states such as diabetes, gout, thyroid disease, hyperlipidemia, obesity, metabolic syndrome, sarcopenia, and cognitive impairment based on the eye image.

[0058] The electronic device (150) can calculate the biological age of the subject by taking into account the condition of the organs, metabolic state, etc.

[0059] The electronic device (150) can generate a customized health guide for the subject by taking into account the subject's biological age, chronological age, organ condition, metabolic condition, etc.

[0060] In this way, an electronic device (150) according to one embodiment of the present invention can predict the whole-body biological age of a subject based on an eye image. The electronic device (150) can receive at least one eye image as input, process it to analyze the condition of at least one organ, and calculate the whole-body biological age of the subject. The electronic device (150) can utilize an artificial intelligence model designed around a neural network-based deep learning structure to characterize the condition of each organ using data extracted from the eye image, and perform the process of calculating the whole-body biological age based thereon.

[0061] FIG. 2 is a block diagram showing the configuration of an electronic device for predicting biological age according to one embodiment of the present invention.

[0062] As illustrated in FIG. 2, the electronic device (150) may include a bus (210), a display (220), a communication circuit (230), a database (240), a memory (250), an I / O interface (260), and a processor (270). In other embodiments, the electronic device (150) may omit at least one of the above components or additionally include other components.

[0063] For reference, the components (210, 220, 230, 240, 250, 260, 270) of the electronic device (150) illustrated in FIG. 2 are merely exemplary components for explaining a method for predicting biological age according to an embodiment of the present invention. That is, it is evident that the electronic device (150) according to an embodiment of the present invention may additionally include other components other than those illustrated.

[0064] The bus (210) can electrically connect the components (220 to 270) to each other. The bus (210) may include circuits for communication (e.g., control messages and / or data) between the components (220 to 270).

[0065] The display (220) can display text, images, videos, icons, or symbols that constitute various content. The display (220) may include a touchscreen and can receive touch, gesture, proximity, or hovering input using an electronic pen or a part of the user's body.

[0066] For example, the display (220) may include a liquid crystal display (LCD), a light-emitting diode (LED) display, an organic light-emitting diode (organic LED) display, a microelectromechanical systems (MEMS) display, or an electronic paper display. The display (220) may be implemented by being included in the electronic device (150), or implemented separately from the electronic device (150) but operatively connected to the electronic device (150).

[0067] The communication circuit (230) can establish a communication channel between the electronic device (150) and external devices. The communication circuit (230) can communicate with external devices by accessing the network (280) via wireless or wired communication. For example, the communication circuit (230) can transmit and receive necessary data with the terminal (110). More specifically, the electronic device (150) can receive eye images of the subject, user data, etc. from the terminal (110) through the communication circuit (230). The electronic device (150) can transmit various data, including health guides, to the terminal (110) through the communication circuit (230).

[0068] The database (240) may be implemented in memory (250) or on a separate storage medium. The database (240) may store all contents, history, etc. of data transmitted and received with the terminal (110). Data stored in the database (240) may be updated regularly according to a predetermined period, and may be updated frequently when new data is input through the terminal (110).

[0069] According to an embodiment of the present invention, various information received from a terminal (110) may be stored in the database (240). For example, the database (240) may store data such as the subject's eye images, health data, electronic health records (EHR) data, and medical imaging records (PACS - Picture Archiving and Communication System) data. For example, the database (240) may operate in a cloud-based manner.

[0070] According to various embodiments, since the data stored in the database (240) is sensitive information of the user, it may be distributed and stored in a blockchain network to enhance security regarding the use of said information. When the database (240) is distributed and stored in a blockchain network, the history of transmission, modification, deletion, addition, etc. of the information contained in the database (240) can be managed more securely in said blockchain network.

[0071] The memory (250) may include volatile and / or non-volatile memory. The memory (250) may store instructions or data related to at least one other component of the electronic device (150). For example, the memory (250) may store instructions that cause the processor (270) to perform various operations described herein at runtime. For example, the instructions may be included in a package file of an application program.

[0072] The I / O interface (260) can perform the role of transmitting commands or data input from a user or other external device to other components of the electronic device (150). The I / O interface (260) can be implemented in hardware or software and can be used as a concept encompassing a user interface (UI) and a terminal for communication with other external devices.

[0073] The processor (270) may include at least one of a central processing unit (CPU), an application processor (AP), or a communication processor (CP). The processor (270) is electrically connected to a memory (250), a display (220), and a communication circuit (230) via a bus (210), and during operation, it may execute operations or data processing regarding the control and / or communication of other components according to instructions, programs, or software stored in the memory (250). Accordingly, the execution of the instructions, application programs, or software can be understood as the operation of the processor (270).

[0074] The processor (270) may receive an eye image of the subject from the terminal (110). According to an embodiment of the present invention, the eye image may include an image of the anterior eye and an image of the posterior eye. According to an embodiment of the present invention, the eye image may include a fundus image. The processor (270) may receive health data of the subject from the terminal (110). According to an embodiment of the present invention, the health data may include data that can affect the subject's health status, such as blood pressure, heart rate, blood oxygen saturation, respiratory rate, number of steps, calories burned, type and time of exercise, distance traveled, sleep duration and quality, sleep stage, movement during sleep, menstrual cycle, stress level, lung capacity, weight, body fat percentage, muscle mass, medication intake, prescription of digital therapeutics, prescription of electronic drugs, comprehensive lifestyle intervention, etc. According to an embodiment of the present invention, the processor (270) may receive electronic health record data from a local server or a remote server within a medical institution. According to an embodiment of the present invention, electronic health record data may include information regarding demographic data, medical history, family medical history, vaccination records, treatment records, prescriptions, diagnostic results, medical staff records, surgery records, medication records, etc.

[0075] The processor (270) can determine the organ condition of the subject by analyzing the eye image provided by the terminal (110). The organ condition may include one or more conditions related to the heart, kidneys, brain, lungs, and liver. For example, the processor (270) can extract the risk level of at least one organ. According to one embodiment of the present invention, the processor (270) can extract the risk level of the organ based on the eye image by utilizing an artificial intelligence model that has already been established.

[0076] The processor (270) can calculate the age of a corresponding organ based on the condition of at least one organ. According to one embodiment of the present invention, the processor (270) can calculate the age of an organ based on the risk level of the organ by utilizing an artificial intelligence model that has already been established.

[0077] The processor (270) can predict the biological age of the subject based on the calculated age of the organ. According to one embodiment of the present invention, the processor (270) can predict the biological age of the subject based on the age of the organ by utilizing an artificial intelligence model that has already been established.

[0078] The processor (270) can generate a health guide corresponding to the subject by comprehensively considering the subject's biological age, chronological age, and organ condition. According to one embodiment of the present invention, the processor (270) can generate a health guide corresponding to the subject by comprehensively considering the subject's biological age, chronological age, and organ condition using an already established artificial intelligence model.

[0079] The network (280) may include at least one of a telecommunications network, a computer network, the Internet, or a telephone network. A wireless communication protocol for accessing the network (280) may use, for example, at least one of LTE (Long-Term Evolution), LTE-A (LTE Advanced), CDMA (Code Division Multiple Access), WCDMA (Wideband CDMA), UMTS (Universal Mobile Telecommunications System), WiBro (Wireless Broadband), GSM (Global System for Mobile communications), or 5G standard communication protocols. However, this is exemplary, and various wired and wireless communication technologies applicable in the relevant technical field may be used according to the embodiments to which the present invention is applied.

[0080]

[0081] FIG. 3 is a conceptual diagram showing the operation of an electronic device (150) for predicting biological age according to one embodiment of the present invention. In particular, FIG. 3 schematically shows the structure and operation of an electronic device (150) for predicting the biological age of a subject using a plurality of artificial intelligence models.

[0082] The electronic device (150) can predict the biological age of a subject through a processor (270) equipped with a pre-learned age prediction model. According to one embodiment of the present invention, the electronic device (150) may include a data processing unit equipped with an age prediction model.

[0083] An age prediction model according to an embodiment of the present invention may include a first model (310) and a second model (320).

[0084] The first model (310) can extract key features related to the state of at least one organ from an eye image provided to the electronic device (150) and predict the health status (risk level) of the organ. According to one embodiment of the present invention, the first model (310) can extract information such as the blood vessel pattern, the thickness of the nerve fiber layer, and the tissue structure of the retinal image, and characterize the health status of at least one organ based on this. In this process, features reflecting the physiological state related to major organs such as the heart, kidneys, and brain are learned, and the risk level for each organ can be calculated. Through this, the electronic device (150) can comprehensively evaluate the health status of the subject using an age prediction model and detect abnormalities in specific organs at an early stage.

[0085] The second model (330) can calculate the age of the organ based on the result data of the first model (310) and calculate the subject's whole-body biological age from this. According to one embodiment of the present invention, the second model (330) can predict the individual age of the organ by reflecting the degree of aging and health status according to the condition of the organ. The second model (330) can calculate the subject's whole-body biological age by utilizing a weighting algorithm on the predicted individual ages of the organs. Each weight applied to each organ can be set to reflect the influence of the organ's health status on the subject's overall biological age. Through this method, the electronic device (150) can effectively predict the subject's whole-body biological age by reflecting the contribution of organs important to health and aging, rather than simply summing the conditions of each organ.

[0086] According to an embodiment of the present invention, the first model (310) and the second model (330) have a structure that integrates the characteristics of input data and the general body condition, thereby enabling personalized health assessment and early detection of general body diseases. The general body condition may include one or more conditions related to obesity, diabetes, dyslipidemia, sarcopenia, dyspnea, cognitive impairment, and hypertension. An age prediction model including the first model (310) and the second model (330) can calculate the general body biological age by precisely analyzing the conditions of multiple organs.

[0087] According to an embodiment of the present invention, the first model (310) and the second model (330) can be trained by adding at least one of an anterior segment image, other measurement data, or demographic data to an input image (e.g., fundus image, retinal image, etc.).

[0088]

[0089] FIG. 4 is a flowchart illustrating the operation of an electronic device (150) for predicting biological age according to an embodiment of the present invention. In particular, FIG. 4 illustrates the operation of the electronic device (150) predicting the biological age of a subject by utilizing an eye image as input data for a pre-established age prediction model.

[0090] In step S401, the electronic device (150) may receive an eye image from the terminal (110). According to an embodiment of the present invention, the eye image may include an image of the anterior part of the eye and an image of the posterior part of the eye. According to various embodiments of the present invention, the electronic device (150) may receive additional personal data. The personal data may include at least one of anterior part images, health data, or demographic data.

[0091] In step S403, the electronic device (150) can analyze an eye image. According to an embodiment of the present invention, the electronic device (150) can utilize the first model (310) to extract information such as the blood vessel pattern, the thickness of the nerve fiber layer, and the tissue structure of the retinal image.

[0092] In step S405, the electronic device (150) can extract the condition (risk level) of at least one organ based on the analyzed content. The electronic device (150) can calculate the risk level for each organ based on the physiological condition related to major organs such as the heart, kidneys, and brain by utilizing the first model (310).

[0093] In step S407, the electronic device (150) can calculate the organ age based on the risk level. According to an embodiment of the present invention, the electronic device (150) can predict the individual age of the organ by utilizing the second model (330) to reflect the degree of aging and health status according to the condition of the organ.

[0094] In step S409, the electronic device (150) can predict the biological age of the subject by considering the organ age. The electronic device (150) can calculate the subject's whole-body biological age by utilizing the second model (330) and the individual ages of the predicted organs using a weighting algorithm.

[0095]

[0096] FIGS. 5a and FIGS. 5b are conceptual diagrams illustrating the operation of an electronic device (150) for predicting biological age according to one embodiment of the present invention.

[0097] FIG. 5a schematically illustrates the configuration and operation of the first model (310) according to an embodiment of the present invention.

[0098] The first model (310) can be designed as a neural network-based structure that extracts key features related to the condition of at least one organ of a subject based on an eye image and predicts the risk level for the condition of each organ.

[0099] According to an embodiment of the present invention, the first model (310) may include a backbone module (311) and individual task modules (313).

[0100] The backbone module (311) may be a common neural network module that extracts key features from an eye image.

[0101] According to an embodiment of the present invention, the first model (310) can learn important information (feature information), such as retinal blood vessel patterns, tissue thickness, and nerve fiber layer damage, from high-resolution eye images through the backbone module (311), and convert this into a multidimensional feature vector.

[0102] The backbone module (311) can be designed based on a Convolutional Neural Network (CNN) and can process input data through a multilayer filter to extract visual features of the image (vascular structure, tissue density, etc.), reduce the complexity of the data, and extract representative features required for organ-specific analysis into high-dimensional vectors.

[0103] The feature vector extracted from the backbone module (311) can be provided to individual task modules (313) that predict the state of each organ.

[0104] The individual task module (313) can independently derive analysis results specialized for at least one organ to subdivide the status of each organ and finally output the risk level of the organ.

[0105] In FIG. 5a, the individual task module (313) is depicted as one, but it can be composed of individual task modules corresponding to each of the multiple organs, i.e., multiple individual task modules. Through this, the first model (310) can have a multi-test structure that effectively processes input data and analyzes the state of each organ. That is, the first model (310) can be operated in a multi-test manner that simultaneously analyzes the state of multiple organs based on the backbone module (311) and the individual task module (313).

[0106] Individual task modules (313) can learn unique information related to the condition of a specific organ and can perform different learning for each specific organ.

[0107] According to an embodiment of the present invention, an individual task module for extracting cardiac risk can learn retinal vascular features (e.g., shape and structural pattern of retinal blood vessels, retinal artery stenosis and changes in blood vessel wall thickness, blood vessel branching angle and asymmetry, etc.) and features associated with cardiovascular indicators such as CAC (Coronary Artery Calcium).

[0108] According to an embodiment of the present invention, an individual task module for extracting kidney risk can learn features affecting kidney function, such as retinal microvascular or blood flow abnormalities related to eGFR (e.g., retinal hemorrhage, retinal microvascular abnormalities, irregularity of vascular branching structure, micro-blood flow changes such as decreased retinal capillary density, changes in the thickness of the retinal choroid, etc.).

[0109] According to an embodiment of the present invention, an individual task module for extracting brain risk can learn features that affect brain risk, such as retinal structure and vascular changes associated with amyloid beta accumulation (retinal nerve fiber layer thickness, macula structure changes, asymmetry and low-density blood flow of retinal blood vessels, abnormal findings around the optic nerve, etc.).

[0110] Each individual task module utilizes an independent loss function to optimize prediction accuracy related to a specific organ, thereby enabling the provision of segmented risk levels for each organ.

[0111]

[0112] FIG. 5b schematically illustrates the configuration and operation of a second model (330) according to an embodiment of the present invention.

[0113] The second model (330) can predict the age of individual organs and the whole-body biological age of the subject based on the individual risk of at least one organ provided by the first model (310). The second model (330) includes statistical methodologies and machine learning-based regression analysis and can process multivariate data integrally to predict the whole-body biological age.

[0114] According to one embodiment of the present invention, the second model (330) may include a long-term age prediction module (331) and an integrated age prediction module (333).

[0115] According to one embodiment of the present invention, the organ age prediction module (331) can perform regression analysis by setting the organ-specific risk input into the second model (330) as an independent variable and the organ-specific biological age as a dependent variable. In this way, the organ age prediction module (331) can calculate the individual biological age of each organ, such as the heart, kidney, brain, lungs, and liver. The regression analysis process can be performed using linear regression, logistic regression, or more complex machine learning models (e.g., Random Forest, LASSO, etc.).

[0116] According to one embodiment of the present invention, the organ age prediction module (331) can calculate the age of each organ by utilizing a specific biological indicator-based or risk assessment model-based method.

[0117] According to one embodiment of the present invention, a long-term age prediction module for predicting the age of the heart may calculate the biological age of the heart based on Coronary Artery Calcium (CAC) scores and retinal blood vessel density data, or by utilizing risk assessment results related to Cardiovascular Disease (CVD). For example, the long-term age prediction module for predicting the age of the heart may quantify an individual's risk using cardiovascular risk assessment models such as CAC scores, CVD risk scores, Pooled Cohort Equations (PCE), Systematic Coronary Risk Evaluation 2 (SCORE2), Predicting Risk of Cardiovascular Disease Events (PREVENT), Framingham Risk Score (FRS), or QRISK, and then convert this into biological age. As the CAC score increases or the CVD risk score increases, the biological age of the heart may be calculated to be higher than the chronological age.

[0118] According to one embodiment of the present invention, a long-term age prediction module for predicting the age of the kidney can calculate the age of the kidney based on eGFR (estimated glomerular filtration rate using the CKD-EPI formula or MDRD formula), creatinine levels, blood urea nitrogen levels, urinary creatinine levels, urinary protein levels, or other kidney function indicators. The long-term age prediction module for predicting the age of the kidney can quantify the kidney function status using chronic renal failure risk assessment indicators (e.g., the Kidney Failure Risk Equation developed in Canada) and then convert this into biological age. If the eGFR value is low or the decline in kidney function is severe, the biological age of the kidney may be calculated to be higher than the chronological age.

[0119] According to one embodiment of the present invention, a long-term age prediction module for predicting brain age can calculate the age of the brain based on indicators related to neurodegenerative diseases such as Alzheimer's disease and dementia. The long-term age prediction module for predicting brain age can quantify the degree of brain aging by utilizing brain disease indicators such as the CAIDE Dementia Risk Score or Amyloid Beta levels. In addition, the long-term age prediction module for predicting brain age can calculate the biological age of the brain by analyzing the thickness of the neurofibrillary layer and the state of vascular abnormalities together. For example, if the Amyloid Beta level is high or the thickness of the neurofibrillary layer is reduced, the biological age of the brain may be evaluated as higher than the chronological age.

[0120] Each organ age calculated in this manner can be provided to the integrated age prediction module (333). The integrated age prediction module (333) can calculate an individual's whole-body biological age by integrating through a weighted algorithm. The weighted algorithm reflects the relative importance of each organ to the whole-body health status, thereby enabling a more accurate prediction of the whole-body biological age.

[0121] Whole-body biological age can be calculated not only using a simple weighted average method but also by utilizing multivariate regression models that reflect the interactions between individual organs. For example, by integrally considering the correlation between heart and kidney conditions, the overall state of whole-body aging can be assessed more precisely.

[0122] In this process, the second model (330) can relatively assess the user's whole-body biological age by comparing the average risk score by age with the individual's risk score. Through this, the health risk can be visualized by comparing the user's health status with that of the same age group, and directions for improving age-related health status can be suggested. This approach is effective in supporting personalized health assessment and preventive health care.

[0123]

[0124] FIGS. 6a and 6b are conceptual diagrams illustrating the operation of an electronic device (150) for predicting biological age according to an embodiment of the present invention. In particular, FIGS. 6a and 6b are conceptual diagrams schematically illustrating the process of an electronic device (150) that receives a fundus image as input data and performs image analysis, organ risk extraction, organ age calculation, and biological age prediction of a subject.

[0125] Referring to FIG. 6a, the electronic device (150) can receive a fundus image and input it into the first model (310). The first model (310) can learn important information such as retinal blood vessel patterns, tissue thickness, and nerve fiber layer damage from the fundus image using the backbone module (610), and convert this into a multidimensional feature vector. The backbone module (610) can extract representative features necessary for organ-specific analysis into a vector and provide the extracted feature vector to the first to third individual task modules (631 to 633) that predict the state of the first organ, the second organ, and the third organ, respectively.

[0126] The first individual task module (631) can extract the condition (risk level) of an organ based on a feature vector corresponding to the first organ provided by the backbone module (610). For example, if the first organ is the heart, the first individual task module (631) can extract the risk level of the first organ based on features associated with retinal blood vessel patterns and cardiovascular indicators such as CAC (Coronary Artery Calcium).

[0127] The second individual task module (632) can extract the condition (risk level) of the organ based on a feature vector corresponding to the second organ provided by the backbone module (610). For example, if the second organ is the kidney, the second individual task module (632) can extract the risk level of the second organ based on features affecting kidney function, such as vascular damage and eGFR.

[0128] The third individual task module (633) can extract the condition (risk level) of the organ based on a feature vector corresponding to the third organ provided by the backbone module (610). For example, if the third organ is the brain, the third individual task module (633) can extract the risk level of the third organ based on the thickness of the nerve fiber layer and features related to amyloid beta.

[0129] Referring to FIG. 6b, the risk level of each of the first to third organs extracted from the first model (310) can be input into the organ age prediction modules (651 to 655) corresponding to each.

[0130] The first organ age prediction module (651) can calculate the biological age of the first organ based on the first organ risk. For example, the first organ age prediction module (651) can perform regression analysis by setting the first organ risk as an independent variable and the biological age of the first organ as a dependent variable. As another example, if the first organ is the heart, the first organ age prediction module (651) can calculate the biological age of the heart based on Coronary Artery Calcium (CAC) scores and retinal blood vessel density data, or by utilizing risk assessment results related to Cardiovascular Disease (CVD).

[0131] The second organ age prediction module (652) can calculate the biological age of the second organ based on the second organ risk. For example, the second organ age prediction module (652) can perform regression analysis by setting the second organ risk as an independent variable and the biological age of the second organ as a dependent variable. As another example, if the second organ is the kidney, the second organ age prediction module (652) can calculate the age of the kidney based on eGFR (estimated glomerular filtration rate using the CKD-EPI formula or MDRD formula), creatinine levels, blood urea nitrogen levels, urine creatinine levels, urine protein levels, or other kidney function indicators. The organ age prediction module for predicting the age of the kidney can quantify the kidney function status using chronic renal failure risk assessment indicators (e.g., the Kidney Failure Risk Equation developed in Canada) and then convert this into a biological age.

[0132] The third organ age prediction module (653) can calculate the biological age of the third organ based on the third organ risk. For example, the third organ age prediction module (653) can perform regression analysis by setting the third organ risk as an independent variable and the biological age of the third organ as a dependent variable. As another example, if the third organ is the brain, the third organ age prediction module (653) can calculate the age of the brain based on indicators related to neurodegenerative diseases such as Alzheimer's disease and dementia. The organ age prediction module for predicting the age of the brain can quantify the degree of brain aging by utilizing brain disease indicators such as the CAIDE Dementia Risk Score or Amyloid Beta levels. Additionally, the organ age prediction module for predicting the age of the brain can calculate the biological age of the brain by analyzing the thickness of the neurofibrillary layer and the abnormal state of blood vessels together.

[0133] The biological ages of the first to third organs calculated from each of the first to third organ age prediction modules (651 to 653) can be provided to the integrated age prediction module (670).

[0134] The integrated age prediction module (670) can apply weights to the biological ages of the first to third organs and integrate them through a weighted algorithm to calculate the individual's whole-body biological age.

[0135] According to an embodiment of the present invention, the integrated age prediction module (670) can calculate the whole-body biological age using the following mathematical formula.

[0136] [Mathematical Formula 1]

[0137]

[0138] Each weight (b1, b2, b3) can be set to reflect the impact of organ health status on an individual's whole-body biological age. This approach allows for a more comprehensive representation of an individual's whole-body biological age by reflecting the greater contribution of organs important for health and aging, rather than simply summing the status of individual organs.

[0139] FIG. 6c is a conceptual diagram illustrating an example of analysis information regarding the whole-body biological age of a subject according to an embodiment of the present invention. In particular, FIG. 6c illustrates an example of information analyzed by integrating multiple organ ages.

[0140] According to an embodiment of the present invention, the electronic device (150) can analyze the subject's whole-body biological age and organ-specific age by utilizing an artificial neural network model.

[0141] According to one embodiment of the present invention, the electronic device (150) can calculate the position of the subject's whole-body biological age within a normal distribution created by collecting the whole-body biological ages of subjects of the same age group.

[0142] According to one embodiment of the present invention, the electronic device (150) can calculate the age range for each organ and output it to the subject through a display. In addition, the electronic device (150) can provide the subject with risk levels and specific characteristics for each organ.

[0143] According to one embodiment of the present invention, the electronic device (150) can provide recommendations based on the results of an age analysis by organ, a condition analysis, and whole-body biological age.

[0144] FIGS. 7a and 7b are conceptual diagrams illustrating the operation of an electronic device (150) for predicting biological age according to an embodiment of the present invention. In particular, FIGS. 7a and 7b are conceptual diagrams schematically illustrating the process of an electronic device (150) that receives a fundus image as input data and performs image analysis, organ risk extraction, metabolic risk extraction, organ age calculation, metabolic age calculation, and prediction of the biological age of a subject.

[0145] Referring to FIG. 7a, the electronic device (150) can receive a fundus image and input it into the first model (310). The first model (310) can learn important information such as retinal blood vessel patterns, tissue thickness, and nerve fiber layer damage from the fundus image using the backbone module (710), and convert this into a multidimensional feature vector. The backbone module (710) can extract representative features necessary for organ-specific analysis into a vector and provide the extracted feature vector to the first and second individual task modules (731 to 732) that predict the state of the first organ and the second organ, respectively.

[0146] Meanwhile, the backbone module (710) can extract representative features necessary for metabolic analysis into a vector and provide the extracted feature vector to a third individual task module (733) that predicts the metabolic state.

[0147] The first individual task module (731) can extract the state (risk level) of the organ based on the feature vector corresponding to the first organ provided by the backbone module (710).

[0148] The second individual task module (732) can extract the state (risk level) of the organ based on the feature vector corresponding to the second organ provided by the backbone module (710).

[0149] The third individual task module (733) can extract the metabolic state (risk level) based on the feature vector corresponding to the metabolic analysis provided by the backbone module (710).

[0150] Referring to FIG. 7b, the risk of each of the first and second organs extracted from the first model (310) can be input into the corresponding organ age prediction modules (751 and 752).

[0151] The first organ age prediction module (751) can calculate the biological age of the first organ based on the first organ risk. The second organ age prediction module (752) can calculate the biological age of the second organ based on the second organ risk.

[0152] Meanwhile, the metabolic age prediction module (753) can calculate the metabolic age based on the metabolic risk level.

[0153] The biological age of the first and second organs calculated from each of the first and second organ age prediction modules (751 and 752) and the metabolic age calculated from the metabolic age prediction module (753) can be provided to the integrated age prediction module (770).

[0154] The integrated age prediction module (770) can predict the biological age of the subject by considering the biological age and metabolic age of the first and second organs.

[0155] The integrated age prediction module (770) can apply weights to the biological age and metabolic age of the first and second organs and integrate them through a weighted algorithm to calculate the individual's whole-body biological age.

[0156] According to an embodiment of the present invention, the integrated age prediction module (770) can calculate the whole-body biological age using the following mathematical formula.

[0157] [Mathematical Formula 2]

[0158]

[0159] Each weight (b1, b2, b3) can be set to reflect the influence of organ health status and metabolic age on an individual's whole-body biological age. This approach allows whole-body biological age to be expressed holistically by reflecting the greater contribution of organs important for health and aging, rather than simply summing the status and metabolic age of each organ.

[0160] That is, the electronic device (150) can process data representing organ status and data representing metabolic status in parallel, and combine the data at the final layer to generate a result. For example, the electronic device (150) can evaluate a health status related to cardiovascular disease by combining vascular status and CRP levels (inflammation levels) extracted from fundus images. As another example, the electronic device (150) can reflect the correlation between kidney status and metabolic status in a model by combining eGFR (kidney function) and BMI data.

[0161] The structure of such an electronic device (150) not only analyzes multiple organ state data but also provides the advantage of being able to evaluate the subject's health condition more precisely and comprehensively by utilizing metabolic state data in an integrated manner.

[0162] FIG. 7c is a conceptual diagram illustrating an example of analysis information regarding the whole-body biological age of a subject according to an embodiment of the present invention. In particular, FIG. 7c illustrates an example of information analyzed by integrating organ age and metabolic age.

[0163] According to an embodiment of the present invention, the electronic device (150) can analyze the subject's whole-body biological age, organ-specific age, and metabolic age by utilizing an artificial neural network model.

[0164] According to one embodiment of the present invention, the electronic device (150) can calculate the position of the subject's whole-body biological age within a normal distribution created by collecting the whole-body biological ages of subjects of the same age group.

[0165] According to one embodiment of the present invention, the electronic device (150) can calculate the age range for each organ and output it to the subject through a display. In addition, the electronic device (150) can provide the subject with risk levels and specific characteristics for each organ.

[0166] According to one embodiment of the present invention, the electronic device (150) can calculate the metabolic age and output it to the subject through a display.

[0167] According to one embodiment of the present invention, the electronic device (150) can provide recommendations based on the results of organ-specific age analysis, metabolic age analysis, organ condition analysis, metabolic condition analysis, and whole-body biological age.

[0168] FIGS. 8A and 8B are conceptual diagrams illustrating the operation of an electronic device (150) for predicting biological age according to an embodiment of the present invention. In particular, FIGS. 8A and 8B illustrate an embodiment of the operation method of an electronic device (150) for converting a cardiovascular risk score into a cardiac biological age using a data-based approach.

[0169] First, obtain data on the distribution of risk scores by age group.

[0170] Referring to Fig. 8a, for the user marked with an asterisk in Fig. 8a, the actual age is between 35 and 39 and the risk score is 17. At this time, the difference between the median for each age group and the user's risk score (|R_user - R_50|) is calculated, and the age group with the minimum difference is selected.

[0171] In this process, the user falls into the 30s to 34s age group. This means that the user has a younger cardiac biological age compared to their actual age.

[0172] In specific situations, heart age can be finely adjusted by considering relationships of different quantiles instead of the median, or by applying weights to risk scores.

[0173] A similar method can be used to calculate the age of other organs, such as the kidneys and brain.

[0174] Referring to Figure 8b, the risk level of the subject based on the predicted whole-body biological age can be determined. Data on whole-body biological age measured from the same age group is collected, and the subject's position within the normal distribution can be identified by comparing the subject's whole-body biological age.

[0175] As shown in Fig. 8b, the median (50%), 25%, 75%, and 90% percentile data for each age group may be included. This defines the trend and distribution of risk scores by age group.

[0176] Compare the user's risk score (R_user) with the median for each age group (R_50) to find the age group where the user's risk score is closest to the median.

[0177] FIG. 9 is a conceptual diagram showing the operation of an electronic device (150) for predicting biological age according to one embodiment of the present invention.

[0178] According to one embodiment of the present invention, the electronic device (150) can receive eye images, demographic data, and other biological data and pre-train an age prediction model through a multimodal method. This method enables more precise whole-body biological age prediction and health status evaluation by integrally learning various data types.

[0179] A multimodal learning method according to one embodiment of the present invention can provide high prediction accuracy by integrally processing various data sources compared to a single data input method. This method allows a neural network model to derive analysis results in various forms, such as processing multiple types of data in parallel, processing them simultaneously, or using a multitasking method.

[0180] According to an embodiment of the present invention, the electronic device (150) may additionally be provided with data such as eye images, health data, and electronic record information. The provided data is processed by independent paths for each individual model, and each path may be configured with a neural network architecture specialized for the data type. This structure learns unique features from each data source and enables more precise prediction by integrally analyzing these data.

[0181] For example, by analyzing the condition of retinal blood vessels from fundus images, learning physical features such as corneal thickness through anterior segment images, and combining them with individual demographic data, it is possible to quantitatively predict the risk of cardiovascular or systemic diseases. This multimodal learning approach provides more sophisticated and reliable results compared to single-data input methods, and enables a comprehensive assessment of overall health status by learning the interactions between various data.

[0182] According to one embodiment of the present invention, a new input method for an age prediction model pre-trained by a multimodal method may include a single input method and a multiple input method.

[0183] The single input method according to an embodiment of the present invention predicts organ status and whole-body biological age using only a single eye image as input data. A neural network model learns key features of fundus images (e.g., vascular patterns, nerve fiber layer thickness, tissue density, etc.) to analyze the individual's organ-specific status and calculates whole-body biological age based on this. The single input method is primarily useful for assessing the initial risk of systemic diseases such as cardiovascular disease, kidney disease, and diabetes. It offers the advantages of providing fast and effective analysis with only simple input data and facilitating data collection. However, since it may be difficult to assess multifaceted health status with only single data, a multiple input method is utilized when additional data input is required.

[0184] The multiple input method according to an embodiment of the present invention is designed to enable multimodal learning by including not only fundus images but also additional personal data. The additional data may include one or more of anterior segment images, health data, other measurement data, electronic medical record data, medical imaging record PACS (Picture Archiving and Communication System) data, or demographic data. In this method, each data type is simultaneously transmitted as input to the model, and the neural network processes the features of each data individually and then integrates and analyzes them. For example, fundus images can provide information on the vascular status of the retina and the nerve fiber layer, while anterior segment images can learn additional physical features through data such as corneal thickness and anterior chamber depth. Personal data (e.g., blood pressure, heart rate, BMI, blood test results, etc.) can complementarily analyze risk factors for diseases. The multiple input method enables multidimensional health status assessment and is particularly useful for accurately predicting the risk of specific diseases (e.g., chronic kidney disease, metabolic syndrome, etc.). For example, vascular data from fundus images can be combined with renal function measurements to calculate renal age or more precisely assess cardiovascular risk. This approach improves predictive accuracy compared to single data input and contributes to personalized health assessments and the development of treatment strategies.

[0185] According to various embodiments of the present invention, the calculated whole-body biological age can be applied to an algorithm that predicts the risk of a second specific disease (e.g., Chronic Obstructive Pulmonary Disease, COPD). Even when the retina has not been proven to be a direct biomarker for a specific disease, the risk of the disease can be effectively predicted by utilizing an intermediate indicator called whole-body biological age.

[0186] According to various embodiments of the present invention, the electronic device (150) may introduce various statistical and machine learning approaches to model the relationship between whole-body biological age and specific diseases. Statistical approaches may include multivariate analysis and Cox proportional hazards models. Multivariate analysis can evaluate the correlation between whole-body biological age and disease risk factors. Cox proportional hazards models may be used to model the time-dependent relationship between whole-body biological age and the likelihood of disease occurrence. Additionally, the electronic device (150) may quantitatively predict the risk of various diseases, including COPD, by combining whole-body biological age and clinical data through a deep learning-based prediction model. Deep learning approaches may include a multi-layer perceptron (MLP) or a multi-task learning model capable of learning multiple indications simultaneously.

[0187] The whole-body biological age calculated according to the present invention indirectly reflects the aging status of specific organs (e.g., cardiovascular system, lungs, kidneys, etc.) using some of the components of the eye (e.g., retina) as biomarkers, and thereby can be used as a key variable for predicting the risk of secondary diseases. This approach is applicable even when the retina has not been proven to be a direct biomarker for a specific disease.

[0188] According to an embodiment of the present invention, the electronic device (150) can improve the accuracy of the prediction of the whole-body biological age by utilizing additional clinical data along with the whole-body biological age. The model for predicting the whole-body biological age according to an embodiment of the present invention can comprehensively analyze the patient's clinical information related to the second disease, such as smoking history, age, BMI, blood pressure, and family history, in addition to the whole-body biological age. Through this, the electronic device (150) can consider the interaction between the whole-body biological age and other clinical variables and can predict disease risk more precisely. For example, the electronic device (150) can analyze COPD risk or evaluate the risk of cardiovascular disease by combining whole-body biological age, smoking history, and BMI.

[0189] The electronic device (150) calculates whole-body biological age based on features extracted from an eye image (e.g., retinal data), and can thereby extend to multiple indications. This approach can significantly expand the scope of application of retinal-based medical AI solutions. In addition, since whole-body biological age can be calculated solely through non-invasive retinal imaging, it reduces the burden on patients and enables early prediction and preventive medical intervention.

[0190] As such, according to an electronic device (150) according to one embodiment of the present invention, the condition of a subject's multiple major organs can be predicted based on an eye image, and the subject's whole-body biological age can be calculated based on the predicted multiple organ conditions. The present invention utilizes the non-invasive nature of the eye image to comprehensively evaluate the subject's organ and whole-body condition, and based on this, can calculate a more precise whole-body biological age. Based on this, a comprehensive health assessment including the subject's whole-body metabolic and circulatory functions can be provided to the subject. As a result, a precise customized health management plan can be established based on the health condition of each organ, and specific guidelines for early disease prediction and preventive medical measures can be provided to the subject, thereby contributing to the improvement of the quality of medical services.

[0191] Although all components constituting an embodiment of the present invention have been described above as being combined or operating in combination, the present invention is not necessarily limited to such embodiments. That is, within the scope of the purpose of the present invention, all components may be selectively combined in one or more ways to operate.

[0192] Meanwhile, the various embodiments described herein may be implemented by hardware, middleware, microcode, software and / or combinations thereof. For example, the various embodiments may be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, other electronic units designed to perform the functions presented herein, or combinations thereof.

[0193] Additionally, for example, various embodiments may be stored or encoded on a computer-readable medium containing instructions. Instructions stored or encoded on a computer-readable medium may enable a programmable processor or other processor to perform a method, for example, when the instructions are executed. A computer-readable medium includes a computer storage medium, and the computer storage medium may be any available medium accessible by a computer. For example, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage media, magnetic disk storage media or other magnetic storage devices.

[0194] Such hardware, software, firmware, etc., may be implemented within the same device or in individual devices to support the various operations and functions described in this specification. Additionally, components, units, modules, components, etc., described as "parts" in this invention may be implemented together or individually as separate but interoperable logic devices. Descriptions of different features of modules, units, etc., are intended to highlight different functional embodiments and do not necessarily imply that they must be realized by individual hardware or software components. Rather, functions associated with one or more modules or units may be performed by individual hardware or software components or integrated within common or individual hardware or software components.

[0195] Although operations are depicted in a specific order in the drawings, it should not be understood that these operations must be performed in the specific order depicted or in a sequential order to achieve the desired result, or that all depicted operations must be performed. In any environment, multitasking and parallel processing may be advantageous. Furthermore, the distinction of various components in the above-described embodiments should not be understood as requiring such distinction in all embodiments, and it should be understood that the described components may generally be integrated together into a single software product or packaged into multiple software products.

[0196] The electronic device, server, or external device according to the various embodiments of the present document described above may include, for example, at least one of a smartphone, tablet PC, mobile phone, video phone, desktop PC, laptop PC, PDA (personal digital assistant), PMP (portable multimedia player), MP3 player, mobile medical device, camera, or wearable device.

[0197] According to various embodiments, the wearable device may include at least one of an accessory type (e.g., a watch, ring, bracelet, anklet, necklace, glasses, contact lens, or head-mounted device (HMD)), a fabric or clothing integrated type (e.g., electronic clothing), a body-attached type (e.g., a skin pad or tattoo), or a bio-implantable type (e.g., an implantable circuit).

[0198] In some embodiments, the electronic device or external device may be a home appliance. The home appliance may include, for example, at least one of a television, a DVD player (Digital Video Disk player), audio, a refrigerator, an air conditioner, a vacuum cleaner, an oven, a microwave oven, a washing machine, an air purifier, a set-top box, a home automation control panel, a security control panel, a TV box, a game console, an electronic dictionary, an electronic key, a camcorder, or a digital photo frame.

[0199] In another embodiment, the electronic device, external device, and wearable device may include at least one of various medical devices (e.g., various portable medical measuring devices (blood glucose meter, heart rate monitor, blood pressure monitor, or body temperature monitor, etc.), MRA (magnetic resonance angiography), MRI (magnetic resonance imaging), CT (computed tomography), imaging device, or ultrasound device, etc.), navigation device, satellite navigation system (GNSS (Global Navigation Satellite System)), EDR (event data recorder), FDR (flight data recorder), automotive infotainment device, home robot, or Internet of Things device (e.g., light bulb, various sensor, electric or gas meter, sprinkler device, fire alarm, thermostat, street light, exercise equipment, hot water tank, heater, boiler, etc.).

[0200]

[0201] As described above, the best embodiments have been disclosed in the drawings and specification. Specific terms have been used herein, but they are used only for the purpose of describing the invention and are not intended to limit the meaning or the scope of the invention as described in the claims. Therefore, those skilled in the art will understand that various modifications and equivalent alternative embodiments are possible therefrom. Accordingly, the true technical scope of protection of the invention should be determined by the technical spirit of the appended claims.

Claims

1. In an electronic device, Memory; and It includes a processor that predicts biological age, The above processor Acquires an eye image of a subject, processes the eye image using a pre-established artificial neural network model, and generates a prediction result for the subject's whole-body biological age. The above-mentioned whole-body biological age is, An electronic device characterized by being determined based on the condition of multiple organs of an individual related to systemic disease.

2. In Claim 1, The artificial neural network model above A first model that processes input to the above artificial neural network model to extract key features related to multiple organ states and predicts the risk level for each organ state; A second model that predicts the individual biological age of each organ based on the risk level for each organ generated from the first model, and determines the whole-body biological age of the subject by integrating the age prediction results for each organ. An electronic device including 3. In Claim 2, The above first model A backbone module for extracting key features from the above eye image and An individual task module that extracts the risk level of at least one organ from features extracted from the above backbone module An electronic device including 4. In Claim 2, The above second model An organ age prediction module that predicts the biological age of an organ based on the risk of at least one organ, and an integrated age prediction module that calculates the whole-body biological age of the subject using a weighting algorithm based on the biological age of the organ. An electronic device including 5. In Claim 1, The above long-term condition is An electronic device characterized by including at least two of the heart, kidney, brain, lung, and liver conditions.

6. In Claim 2, An electronic device characterized in that the above model input includes additional personal data, and the personal data includes at least one of anterior segment images, health measurement data, or demographic data.

7. In Claim 2, An electronic device characterized by the above artificial neural network model being trained by adding at least one of anterior segment images, health measurement data, or demographic data to an input image.

8. In a system for predicting an individual's whole-body biological age based on fundus images, It includes an age prediction model that receives one or more fundus images as model input, processes the fundus images, and predicts an individual's whole-body biological age. The above age prediction model is, Processing model inputs containing one or more fundus images to generate model outputs that characterize an individual's whole-body biological age, and The above-mentioned whole-body biological age is, A system characterized by being determined based on an individual's organ condition and systemic condition.

9. In Claim 8, The above age prediction model is, A first model that analyzes the above model input using a neural network model to extract key features related to organ and general condition and predicts the risk level for said organ and general condition; A system characterized by including a second model that predicts individual ages for each organ and body condition based on the risk of each organ and body condition generated from a first model, and determines an individual's body biological age by integrating the age prediction results for each condition.

10. In claim 8, The above organ conditions include one or more conditions related to the heart, kidneys, brain, lungs, and liver, and A system characterized in that the above-mentioned general condition includes one or more conditions associated with obesity, diabetes, dyslipidemia, sarcopenia, dyspnea, cognitive impairment, and hypertension.

11. In Claim 8, A system characterized in that the above model input additionally includes personal data, and the personal data includes at least one of anterior segment images, health measurement data, or demographic data.

12. In claim 8, A system characterized in that the above neural network is trained using at least one of an input image, an anterior segment image, health measurement data, or demographic data.

13. In a method of operating an electronic device, Step of acquiring an image of the subject's eye; A step of processing the eye image using a previously established artificial neural network model; and Step of generating a prediction result for the whole-body biological age of the above-mentioned subject Includes, The above-mentioned whole-body biological age is, A method of operation of an electronic device characterized by being determined based on the condition of multiple organs of an individual related to systemic disease.

14. In Claim 13, The step of processing the above eye image is A step of analyzing the input to the artificial neural network model to extract key features related to multiple organ states; A step of predicting the risk for the state of each of the above organs; A step of predicting the individual biological age of each organ based on the above risk level; A step of calculating the whole-body biological age of the subject by integrating the biological age prediction results for each organ mentioned above; A method of operation of an electronic device including 15. In Claim 14, The step of extracting the above key features is Characterized by learning feature information related to retinal blood vessel patterns, tissue thickness, and nerve fiber layer damage in the above ocular image, and converting the feature information into a multidimensional feature vector. Method of operation of an electronic device.

16. In Claim 15, Step of extracting the metabolic state of the subject from the above feature vector A method of operation of an electronic device further comprising 17. In Claim 14, The step of calculating the whole-body biological age of the above-mentioned subject is Calculating the whole-body biological age of the subject using a weighting algorithm based on the biological age of the above organs Method of operation of an electronic device.

Images