Method for analyzing eating behavior on basis of biological signals

Abstract

A method for analyzing eating behavior on the basis of biological signals according to the present invention comprises the steps of: receiving, from a user apparatus, body temperature information, heart rate information, and blood pressure information measured by a body temperature measurement unit, a heart rate measurement unit, and a blood pressure measurement unit provided in a wearable apparatus; generating, from the body temperature information, body temperature derivative information that is a derivative value on a time axis, and body temperature integral information that is an integral value in comparison to a reference body temperature; generating a cardiovascular reactivity index by calculating a change amount ratio between the heart rate information and the blood pressure information; predicting eating behavior, a meal amount, and / or a nutrient composition by inputting, to a machine-learning model, the body temperature derivative information, the body temperature integral information, and the cardiovascular reactivity index; and transmitting the predicted result to the user apparatus or linking the prediction result with medical treatment data of a medical institution.

Description

Biosignal-based feeding behavior analysis method

[0001] The present invention relates to a method and system for analyzing eating behavior based on biosignals, and more specifically, to a method and system for analyzing eating behavior based on biosignals that predicts eating behavior, food intake, and the composition of nutrients consumed by measuring and analyzing biosignals through a wearable device worn or attached to a user's body.

[0002]

[0003] Conventional eating behavior monitoring technology has primarily relied on detecting physical actions such as wrist movements, chewing sounds, and swallowing motions using accelerometers, gyroscopes, microphones, and piezoelectric sensors. This approach has limitations in accurately recognizing behavioral patterns and makes it difficult to precisely determine whether the user is actually eating.

[0004] Existing biosignal measurement and analysis methods had limitations in accurately interpreting diverse biosignal patterns because they relied solely on raw information collected from a single sensor or relied on engineer-centric interpretation of the collected data. Furthermore, there were issues such as the difficulty of distinguishing noise and signal patterns with significant individual variability, as well as the high cost and time required for post-processing and analysis.

[0005] In particular, no studies have been identified to date that detect overeating by utilizing information related to the ratio of heart rate (HR) and blood pressure (BP) or body temperature. Existing studies have only analyzed changes in heart rate, blood pressure, or body temperature individually after a meal, and there have been no cases of utilizing them as a ratio. Similarly, regarding information related to body temperature, there have only been cases of analysis using simple measurements, and there have been no cases of utilizing the degree of increase in body temperature through precise analysis.

[0006]

[0007] The biosignal-based eating behavior analysis method according to the present invention aims to accurately predict eating behavior, food intake, nutrient composition, etc., through the analysis of combinations of various biosignals and provide this to the user.

[0008] The biosignal-based feeding behavior analysis method according to the present invention enables more in-depth analysis by utilizing the derivative and integral values ​​of the biosignal, and in particular, has another purpose of detecting whether the amount of food is insufficient through the △BP / △HR ratio.

[0009] The biosignal-based eating behavior analysis method according to the present invention enables more in-depth analysis by utilizing the derivative and integral values ​​of biosignals, and in particular, has another purpose of detecting whether the amount of food consumed is insufficient through the analysis of the rate of change in basal body temperature and the integral value of the change in body temperature within a preset time range.

[0010] The biosignal-based eating behavior analysis method according to the present invention supports the user's health management through real-time monitoring and AI analysis, and has another objective of detecting health abnormalities such as postprandial hypotension or diabetic autonomic neuropathy.

[0011] Another objective of the biosignal-based feeding behavior analysis method according to the present invention is to provide an automated monitoring service in conjunction with wearable devices such as smartwatches, smart glasses, or smart rings.

[0012] Another objective of the biosignal-based eating behavior analysis method according to the present invention is to detect eating information regarding whether the amount of food consumed is insufficient in real time and to provide health feedback by linking with a diet app or a health management platform.

[0013]

[0014] The objectives of the present invention are not limited to those mentioned above, and other unmentioned objectives will be clearly understood by those skilled in the art from the description below.

[0015]

[0016] A method for analyzing eating behavior based on biosignals according to one aspect of the present invention for solving the above technical problem comprises the steps of: a service server receiving biosignal data from at least one wearable device worn or attached to a user's body or a user device connected to said wearable device; generating processed information from the received biosignal data; inputting the generated processed information into a machine learning model to predict the user's eating behavior; and transmitting the predicted eating behavior information to the user device or linking it with medical data from a medical institution.

[0017] Specifically, in a method of operation of a service server that receives biosignal data from at least one wearable device worn or attached to a user's body and predicts eating behavior, the service server may include the steps of: receiving body temperature information, heart rate information, and blood pressure information measured by a body temperature measuring unit, a heart rate measuring unit, and a blood pressure measuring unit equipped in the wearable device from a user device; generating body temperature derivative information, which is a derivative value on a time axis, and body temperature integral information, which is an integral value relative to a reference body temperature, from the body temperature information; generating a cardiovascular response index by calculating the ratio of the change amount of the heart rate information and blood pressure information; inputting the body temperature derivative information, body temperature integral information, and cardiovascular response index into a machine learning model to predict eating behavior, food quantity, and / or nutrient composition; and transmitting the prediction result to a user device or linking it with medical data from a medical institution.

[0018] Here, the service server may further include the step of generating an edge artificial intelligence unit comprising part or all of the machine learning model and the step of transmitting the generated edge artificial intelligence unit to the user device or the wearable device, and the edge artificial intelligence unit may be configured to process and analyze biosignals collected from the wearable device locally in real time to provide immediate feedback without an internet connection.

[0019] Here, the edge artificial intelligence unit transmitted by the service server is optimized for the limited computing resources of the wearable device and can be optimized to the hardware performance of the device through quantization and model compression techniques.

[0020] Here, the service server may further include the step of receiving an analysis result processed through an edge artificial intelligence unit from the wearable device, performing a cloud-based advanced analysis based thereon, and then generating an updated edge artificial intelligence unit and retransmitting it to the wearable device.

[0021] Here, the service server may further include the step of periodically updating the machine learning model by learning biometric information collected from multiple users and actual eating behavior data.

[0022] Here, the service server includes the step of building a personalized prediction model by creating a user-specific profile of a specific user selected among the plurality of users, and may further include the step of adjusting the model weights to be personalized by comparing the actual eating behavior data and the predicted values ​​during the initial 5 to 10 days.

[0023] Here, the service server may further include the step of generating a user-customized edge artificial intelligence unit based on the user-specific profile and corrected model weights and transmitting it to a user device or wearable device so that personalized eating behavior prediction and analysis can be performed locally.

[0024] Here, the cardiovascular response index is defined as the ratio of the change in blood pressure to the change in heart rate (△BP / △HR), and the service server may further include the step of determining an overeating state when the cardiovascular response index is less than -0.3, generating overeating state information, and transmitting it to a user device.

[0025] Here, the service server may be configured to determine an overeating state when the cardiovascular response index is less than -0.5, generate overeating state information, and transmit it to the user device.

[0026] Here, the service server further includes the step of estimating the calorie intake based on the magnitude of the body temperature integral information, classifying the ratios of protein, carbohydrates, and fats based on the pattern of the body temperature differential information to generate nutrient composition ratio information, and may further include the step of transmitting the estimated calorie amount and nutrient composition ratio information to the user device.

[0027] Here, the service server may further include the step of generating an edge artificial intelligence unit including a lightweight artificial intelligence model for calorie estimation and nutrient composition ratio classification and transmitting it to the wearable device, thereby enabling the wearable device to analyze the intake of calories and nutrient composition ratio in real time and provide immediate feedback even without an internet connection.

[0028] Here, the step of generating the nutrient composition ratio information may be configured such that if the gradient of the differential body temperature information is 0.3℃ / 20 min to 0.7℃ / 20 min or more and the duration is 15 min to 25 min or more, the high-protein diet information is generated and transmitted to a user device, or if the gradient of the differential body temperature information is 0.1℃ / 20 min to 0.3℃ / 20 min or less and the duration is 10 min to less than 15 min, the high-fat diet information is generated and transmitted to a user device.

[0029] Here, the service server may further include the step of determining that there is a measurement error when the correlation coefficient of the measurement values ​​between multiple wearable devices is less than 0.5 to 0.7, transmitting a self-correction command to the device, and receiving the self-correction result to use for improving device performance.

[0030] Here, the service server may further include the step of creating a distributed edge artificial intelligence unit for self-correction and data synchronization among the plurality of wearable devices and transmitting it to each wearable device, thereby enabling the plurality of wearable devices to communicate with each other to autonomously correct and synchronize measurement values.

[0031] Here, the service server may further include the step of analyzing the change patterns over time of the body temperature information, heart rate information, and blood pressure information to distinguish between breakfast, lunch, dinner, and snacks; and the step of generating dietary habit analysis information based on the distinction result and transmitting it to a user device.

[0032] Here, the service server further includes the step of sharing anonymized biometric information and meal pattern data with a medical research institution with the user's consent, and may further include the step of updating the machine learning model by receiving an improved analysis algorithm from the medical research institution.

[0033] Here, the service server may further include the step of generating an updated edge artificial intelligence unit based on an improved analysis algorithm received from the medical research institution and transmitting it to the wearable device so that an analysis reflecting the latest medical research results can be performed locally.

[0034] According to another aspect of the present invention, a service server that analyzes biosignals and predicts eating behavior by communicating with at least one wearable device worn or attached to a user's body and / or a user device connected to said wearable device may be provided, comprising: a communication unit that receives body temperature information, heart rate information, and blood pressure information from said wearable device; a data processing unit that generates body temperature differential information and body temperature integral information from said body temperature information and generates a cardiovascular response index from said heart rate information and blood pressure information; an artificial intelligence analysis unit that receives said body temperature differential information, body temperature integral information, and cardiovascular response index and predicts eating behavior, calorie intake, and nutrient composition ratios; a database that stores a plurality of user data and maintains a customized model for each user; and a result providing unit that transmits the prediction results to the user device or links with a medical institution server.

[0035] Here, the service server further includes an edge artificial intelligence distribution unit that generates an edge artificial intelligence unit including part or all of the artificial intelligence analysis unit and transmits it to the wearable device or the user device, and the edge artificial intelligence unit may be configured to perform real-time local processing and analysis of biosignals by being optimized for the hardware performance of the wearable device or the user device.

[0036] A method for analyzing eating behavior based on biosignals according to another aspect of the present invention for solving the above technical problem comprises: a step in which a wearable device worn or attached to a user’s body collects biometric information; a step in which an edge artificial intelligence unit of the wearable device generates processed information from the collected biometric information; a step in which the edge artificial intelligence unit of the wearable device inputs the generated processed information into a machine learning model to predict the user’s eating behavior, food quantity, and / or nutrient composition; and a step in which the wearable device provides the predicted eating behavior information to the user through an output unit or transmits it to a user device or a service server.

[0037] Specifically, in a method for predicting eating behavior in which a wearable device worn or attached to a user's body collects biometric information and an edge artificial intelligence unit included in the wearable device analyzes the collected biometric information, the method may include: a step in which the wearable device collects body temperature information, heart rate information, and blood pressure information, respectively, through a body temperature measuring unit, a heart rate measuring unit, and a blood pressure measuring unit; a step in which the edge artificial intelligence unit of the wearable device generates body temperature derivative information, which is a derivative value on a time axis, and body temperature integral information, which is an integral value relative to a reference body temperature, from the body temperature information; and a step in which the edge artificial intelligence unit of the wearable device calculates the ratio of the change amount of the heart rate information and blood pressure information to generate a cardiovascular response index.

[0038] Here, the wearable device is equipped with two or more independent devices among a wrist-worn device, an ear-worn device, a chest-worn device, and a glasses-worn device, and the wearable device may further include a step of performing self-correction when the correlation coefficient of the measured values ​​between each device is less than 0.5 to 0.7.

[0039] Here, the body temperature integration information is a cumulative value of the increase in body temperature due to digestion-induced heat generation, and can be generated through the steps of: the wearable device continuously measuring the user's body temperature; the edge artificial intelligence unit of the wearable device calculating the rate of change in body temperature over time from the measured body temperature data; and the edge artificial intelligence unit of the wearable device determining rapid food intake when the calculated rate of change in body temperature exceeds a predetermined threshold.

[0040] Here, the edge artificial intelligence unit of the wearable device may be configured to generate and output fast eating status information when it is confirmed that the body temperature rise rate is 0.3℃ / 10 minutes or more and lasts for 10 minutes or more within a measurement period of 30 to 40 minutes.

[0041] Here, the cardiovascular response index is defined as the ratio of the change in blood pressure to the change in heart rate (△BP / △HR), and the edge artificial intelligence unit of the wearable device may further include the step of determining an overeating state when the cardiovascular response index is less than -0.3 and generating and outputting overeating state information.

[0042] Here, the edge artificial intelligence unit of the wearable device may be configured to determine an overeating state when the cardiovascular response index is less than -0.5 and to generate and output overeating state information.

[0043] Here, by utilizing exercise information collected through an accelerometer equipped in the wearable device, the edge artificial intelligence unit of the wearable device may further include a preprocessing step of distinguishing between exercise-induced body temperature rise and meal-induced body temperature rise.

[0044] Herein, the edge artificial intelligence unit of the wearable device inputs the body temperature differential information, body temperature integral information, and cardiovascular response index into a machine learning model to predict eating behavior, food quantity, and / or nutrient composition; and the wearable device transmits the prediction result to a user device or service server to provide feedback; the machine learning model is equipped with a hybrid structure of Long Short-Term Memory LSTM and Temporal Convolutional Network and may be configured to simultaneously analyze the standard deviation of the body temperature differential information, the coefficient of variation of the cardiovascular response index, and the slope of the body temperature integral information within a time window of 10 to 20 minutes.

[0045] Here, the edge artificial intelligence unit of the wearable device may further include the step of estimating the calorie intake according to the magnitude of the body temperature integral information and classifying the ratio of protein, carbohydrates, and fat according to the pattern of the body temperature differential information.

[0046] Here, the edge artificial intelligence unit of the wearable device may be configured to classify the differential body temperature information as a high-protein diet and generate high-protein diet information and transmit the high-protein diet information to the user device or service server when the gradient of the differential body temperature information is 0.3℃ / 20 min to 0.7℃ / 20 min or more and the duration is 15 min to 25 min or more, and to classify it as a high-fat diet and generate high-fat diet information and transmit the high-fat diet information to the user device or service server when the gradient of the differential body temperature information is 0.1℃ / 20 min to 0.3℃ / 20 min or less and the duration is 10 min to less than 15 min.

[0047] Here, the wearable device further includes the step of receiving and installing an edge artificial intelligence unit from a service server or a user device, and the received edge artificial intelligence unit may be configured to include a lightweight artificial intelligence model optimized for the hardware performance of the wearable device.

[0048] Here, the wearable device may further include the step of transmitting body temperature information, heart rate information, and blood pressure information measured at regular intervals in a 24-hour cycle to the service server, along with the eating behavior prediction results, meal amount estimation results, and nutrient composition ratio analysis results generated by the edge artificial intelligence unit, and receiving an updated edge artificial intelligence unit with improved analysis accuracy from the service server and automatically updating it.

[0049]

[0050] The biosignal-based eating behavior analysis method according to the present invention can accurately predict eating behavior, food intake, nutrient composition, etc., through the analysis of combinations of various biosignals and provide them to the user.

[0051] The biosignal-based feeding behavior analysis method according to the present invention enables more in-depth analysis by utilizing the derivative and integral values ​​of the biosignal, and in particular, can detect whether the amount of food consumed is insufficient through the △BP / △HR ratio.

[0052] The biosignal-based eating behavior analysis method according to the present invention enables more in-depth analysis by utilizing the derivative and integral values ​​of biosignals, and in particular, has another purpose of detecting whether the amount of food consumed is insufficient through the analysis of the rate of change in basal body temperature and the integral value of the change in body temperature within a preset time range.

[0053] The biosignal-based eating behavior analysis method according to the present invention supports the user's health management through real-time monitoring and AI analysis, and can even detect health abnormalities such as postprandial hypotension or diabetic autonomic neuropathy.

[0054] The biosignal-based feeding behavior analysis method according to the present invention can provide an automated monitoring service in conjunction with wearable devices such as smartwatches, smart glasses, or smart rings.

[0055] The biosignal-based eating behavior analysis method according to the present invention can detect eating information regarding whether the amount of food is insufficient in real time and provide health feedback by linking with a diet app or a health management platform.

[0056]

[0057] The effects of the present invention are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below.

[0058]

[0059] Figure 1 is a flowchart of the biosignal-based feeding behavior analysis method of the present invention.

[0060] FIG. 2 is a flowchart of a biosignal-based feeding behavior analysis method according to another embodiment of the present invention.

[0061] Figure 3 is a system configuration diagram of the present invention.

[0062]

[0063] The objectives and effects of the present invention, and the technical configurations for achieving them, will become clear by referring to the embodiments described in detail below in conjunction with the accompanying drawings. In describing the present invention, if it is determined that a detailed description of known functions or configurations may unnecessarily obscure the essence of the invention, such detailed description will be omitted. Furthermore, the terms described below are defined with consideration of their function in the present invention, and these may vary depending on the intentions or practices of the user or operator.

[0064] However, the present invention is not limited to the embodiments disclosed below but may be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. Therefore, such definition should be based on the content throughout this specification.

[0065] Throughout the specification, when a part is described as "including" or "equipped" with a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, terms such as "...unit," "...part," or "...module" as used in the specification refer to a unit that processes at least one function or operation, and this may be implemented in hardware, software, or a combination of hardware and software.

[0066] Meanwhile, in an embodiment of the present invention, each component, functional block, or means may be composed of one or more sub-components, and the electrical, electronic, and mechanical functions performed by each component may be implemented by various known devices or mechanical elements such as electronic circuits, integrated circuits, and ASICs (Application Specific Integrated Circuits), and each may be implemented separately or two or more may be integrated into one.

[0067] Additionally, combinations of each block of the attached block diagram and each step of the flowchart may be executed by computer program instructions. Since these computer program instructions can be loaded into the processors of general-purpose computers, specialized computers, portable notebook computers, network computers, mobile devices such as smartphones, online service providers, or other programmable data processing equipment, the instructions executed through the processors of the computer devices or other programmable data processing equipment create means to perform the functions described in each block of the block diagram or each step of the flowchart described below. Since these computer program instructions may also be stored in memory or computer-readable memory available to a computer device capable of orienting the computer device or other programmable data processing equipment to implement functions in a specific manner, it is also possible to produce a manufactured product containing means of instruction that perform the functions described in each block of the block diagram or each step of the flowchart. Since computer program instructions can be loaded onto a computer device or other programmable data processing equipment, it is also possible to create a process for performing a series of operation steps on a computer device or other programmable data processing equipment and to provide steps for executing the functions described in each block of the block diagram and each step of the flowchart.

[0068] Additionally, each block or each step may represent a module, segment, or part of code containing one or more executable instructions for executing a specified logical function(s). Also, it should be noted that in some alternative embodiments, the functions mentioned in the blocks or steps may occur out of order. For example, two blocks or steps described in succession may actually be performed substantially simultaneously, or the blocks or steps may sometimes be performed in reverse order according to the corresponding function.

[0069] In embodiments of the present invention, the term "user device" or "wearable device" refers to any computing means capable of collecting, reading, processing, manipulating, storing, and displaying data, such as a desktop computer, a laptop computer, a smartphone, a PDA, a mobile phone, a game console, a smartwatch, smart glasses, or a smart ring. In particular, the user device in the embodiments of the present invention is a device capable of executing software written in decipherable code and having the function of displaying and transmitting this to the user. Additionally, if necessary, the software may be stored internally or read from an external source along with data. Furthermore, the term "connected" between the user device and the wearable device may refer to a state in which data is transmitted and received in real time between the two devices via wireless or wired communication, enabling mutual information exchange and control. For example, this applies to cases where a smartphone, a smartwatch, or smart glasses are connected wirelessly via Bluetooth or similar methods to exchange information such as text messages, phone calls, and health data, and operate mutually according to user commands. In other words, being "connected" does not simply mean being physically close, but rather that the two devices recognize each other and are in a communication state capable of synchronizing data or exchanging commands.

[0070] In addition, the terminal in the embodiment of the present invention includes not only the data processing function described above but also functions such as input, output, and storage. To this end, it may include various elements found in general computer devices, such as a CPU, motherboard, graphics card, hard disk, sound card, speaker, keyboard, mouse, monitor, USB, and communication modem, as well as external connection terminals found in wireless smartphone terminals, such as a CPU, motherboard, graphics chip, memory chip, sound engine, speaker, touchpad, USB, communication antenna, and a communication modem capable of implementing communication such as 3G, LTE, LTE-A, WiFi, and Bluetooth. These various elements may implement one or more functions individually, in pairs or together, or by combining parts of the various elements. In the embodiment of the present invention, a device or part thereof indicated by one or more blocks in the drawings or detailed description may refer to various elements included in such user devices that represent one or more functions individually, in pairs or together, or by combining parts of the various elements.

[0071] The term 'Edge AI' used in the present invention refers to a technology that collects, processes, and analyzes data by running an artificial intelligence model on the device where data is generated (edge ​​device) itself, rather than on a central server or cloud environment. Specifically, it refers to an artificial intelligence technology that directly collects biosignals from edge devices, such as wearable devices or user devices, and performs real-time analysis and prediction based on them.

[0072] In the present invention, the 'edge artificial intelligence unit' may be defined as a software module comprising part or all of a machine learning model generated by a service server and transmitted to a user device or a wearable device. The edge artificial intelligence unit is configured to process and analyze biosignals collected from a wearable device in real time in a local environment to provide immediate feedback without an internet connection. The 'edge artificial intelligence unit' is optimized for the limited computing resources of the wearable device and can be provided optimized to the hardware performance of the device through quantization and model compression techniques.

[0073] In addition, the above 'edge artificial intelligence unit' is configured to be customized for the user based on user profiles and corrected model weights, so that personalized eating behavior prediction and analysis can be performed locally.

[0074] The edge AI included in the 'edge AI unit' of the present invention can be designed to enable efficient computation even in a limited hardware environment by including a lightweight AI model for calorie estimation and classification of nutrient composition ratios. By adopting a distributed edge AI structure for self-correction and data synchronization among multiple wearable devices, the devices can communicate with each other to autonomously correct and synchronize measurement values. It may possess a hybrid learning structure in which analysis results from the local area are transmitted to a service server, and the server performs cloud-based advanced analysis based on this, generates an updated edge AI unit, and retransmits it to the device.

[0075] The edge AI unit of the wearable device can directly generate body temperature derivative information, which is a derivative value over the time axis, and body temperature integral information, which is an integral value relative to a reference body temperature, from body temperature information. In addition, the edge AI unit can generate a cardiovascular response index by calculating the ratio of changes in heart rate information and blood pressure information in real time. Based on the generated information, it can be configured to immediately predict eating behavior, food intake, and nutrient composition locally and provide them to the user. The wearable device or a user device connected to the wearable device transmits biometric information measured at regular intervals to a service server along with the analysis results generated by the edge AI unit, and automatically updates by receiving an updated edge AI unit with improved accuracy from the server.

[0076] The service server can generate an updated edge AI unit based on an improved analysis algorithm received from a medical research institution, so that the latest medical research results can be reflected in local analysis. The edge AI defined in this way is used in the biosignal-based eating behavior analysis method of the present invention and can be understood as a means of efficiently providing real-time analysis and feedback for the user's health management and improvement of eating habits.

[0077] Meanwhile, in an embodiment of the present invention, a user device, a wearable device, etc. may have a communication function, and to implement the communication function, various network means such as wired internet, wireless internet, infrared communication, Bluetooth, WCDMA, Wibro, WiFi, LTE, LTE-A, and 5G wired / wireless telephone networks may be provided.

[0078] FIG. 1 is a flowchart of a biosignal-based feeding behavior analysis method of the present invention, FIG. 2 is a flowchart of a biosignal-based feeding behavior analysis method according to another embodiment of the present invention, and FIG. 3 is a system configuration diagram of the present invention. Hereinafter, a biosignal-based feeding behavior analysis method according to an embodiment of the present invention will be described in detail.

[0079] In describing the present invention, bio-information refers to information measured in the user's body, such as heart rate (HR), blood pressure (BP), body temperature (BT), respiratory rate (RR), and oxygen saturation (SpO2).

[0080] In describing the present invention, the eating behavior may be understood to include at least one piece of information regarding the start, duration, and end of the meal.

[0081] In describing the present invention, processed information refers to information generated by differentiating, integrating, or combining collected biological information, and includes differential body temperature information, integral body temperature information, cardiovascular response index, etc.

[0082] Differential body temperature information is generated by differentiating body temperature information with respect to time, and represents the rate of increase in body temperature.

[0083] Body temperature integral information is generated by integrating the portion of body temperature data exceeding the reference temperature over time, and represents the cumulative value of the elevated body temperature.

[0084] The cardiovascular response index is an index representing the ratio of heart rate change information to blood pressure change information, and is used to determine whether overeating has occurred.

[0085] Diet-Induced Thermogenesis (DIT) refers to the rise in body temperature that occurs during the digestion process after food intake.

[0086] The present invention relates to a service server (100) for analyzing biosignals collected from a wearable device (210) to predict a user's eating behavior and to analyze the amount of food and nutrient composition, and a method for analyzing eating behavior based on biosignals. The service server (100) of the present invention includes a communication unit, a data processing unit, an artificial intelligence analysis unit, a database, a result providing unit, and an edge artificial intelligence distribution unit.

[0087] The description of the present invention is based on the definition and calculation method of the Cardiovascular Response Rate (CVR). The Cardiovascular Response Rate is defined as the ratio (△BP / △HR) of the change in blood pressure (△BP) and the change in heart rate (△HR), and is used as a key indicator to quantify the hemodynamic load caused by a meal.

[0088] For example, during a normal meal, the heart rate increases within the range of 4–10 bpm and blood pressure decreases by 0–5 mmHg, indicating that the increased blood flow to the digestive organs is balancedly compensated by the cardiovascular system. Conversely, during overeating, the CVR value appears to be less than -0.3, which suggests that the compensatory mechanism of the autonomic nervous system has reached its limit due to reduced peripheral blood flow caused by splanchnic pooling in the abdominal cavity. Through fusion analysis with differential and integral body temperature information, differential body temperature information (dT / dt) suggests a high-protein diet if it exhibits a rapid upward pattern of 0.3–0.7°C 5–15 minutes after the start of the meal, while integral body temperature information (∫(T-T0)dt) indicates the possibility of overeating when a cumulative increase of 0.5°C or more is observed over 30–40 minutes. In particular, if the CVR value is less than -0.3 and the body temperature integral information is 15-20℃·min or higher, it is comprehensively determined to be a state of high-calorie overeating. To perform personalized correction of the machine learning model, the LSTM-Transformer hybrid model simultaneously analyzes the standard deviation of the body temperature derivative information, the coefficient of variation of the CVR, and the slope of the body temperature integral information within a time window of 10 to 20 minutes, and performs model weight correction reflecting individual metabolic differences using user data from the initial 5 to 10 days.

[0089] In the case of an embodiment of the present invention, when the CVR value is less than -0.3, the state of overeating can be detected in real time with 89% accuracy, and when postprandial hypotension occurs, a blood pressure drop of 20 mmHg or more can be automatically detected, and signs of autonomic nervous system dysfunction appearing as a CVR value of +0.3 or higher in diabetic patients can be detected early. In addition, when the gradient of the differential body temperature information is 0.3 to 0.7°C / 20 min and the CVR value is -0.3 to -0.5, a high-protein diet is classified with 82% accuracy, and when the gradient of the differential body temperature information is less than 0.3°C / 20 min and the CVR value is less than -0.3, a high-fat diet is identified with 75% accuracy, thereby significantly improving the nutrient composition classification performance.

[0090] Below, each component of the service server (100) and its function will be described in detail.

[0091] First, the communication unit is configured to receive body temperature information, heart rate information, and blood pressure information from at least one wearable device (210) worn or attached to the user's body and / or a user device (200) connected to the wearable device (210). The communication unit supports various communication protocols (Bluetooth, WiFi, NFC, 5G, etc.) to ensure compatibility with various wearable devices (210). Additionally, the communication unit includes an error detection and correction algorithm to minimize packet loss that may occur during data transmission. The communication unit includes detailed modules such as a protocol conversion module, a data compression / decompression module, a secure communication module, and a connection management module. The protocol conversion module converts the communication protocols of various wearable devices (210) into the internal data format of the service server (100), and the data compression / decompression module is responsible for processing data compression and decompression for bandwidth efficiency. The secure communication module is configured to perform encryption and authentication processing for the secure transmission of biometric information, and the connection management module is responsible for simultaneous connection and session management with multiple devices. The communication unit can also receive metadata such as the battery status and measurement accuracy of the wearable device (210) and use it to evaluate the reliability of the data.

[0092] The data processing unit is equipped to generate differential body temperature information, integral body temperature information, and a cardiovascular response index based on body temperature, heart rate, and blood pressure information received from the communication unit. The differential body temperature information generation module generates an indicator representing the rate of change of body temperature over time and applies moving average filtering for noise removal. This module calculates differential values ​​using the central difference method and performs statistical filtering to remove outliers. The differential body temperature information serves as an important indicator for detecting changes in body temperature at the start of a meal. The integral body temperature information generation module generates an indicator representing the integral value of body temperature relative to a reference body temperature, sets the reference body temperature, calculates the cumulative integral, and performs time window-based integral value normalization. The integral body temperature information is used to measure the cumulative effect of body temperature increase caused by a meal. The cardiovascular response index generation module generates an indicator defined as the ratio of changes in blood pressure and heart rate; it calculates the changes in blood pressure and heart rate relative to a reference state, calculates this ratio, applies time weights, and performs normalization. The cardiovascular response index is used as an important indicator to determine the amount of food consumed and whether overeating has occurred. The data processing unit also includes preprocessing functions such as missing value interpolation, time synchronization, and feature extraction.

[0093] The AI ​​analysis unit receives differential body temperature information, integral body temperature information, and cardiovascular response index generated by the data processing unit as input and performs the role of predicting eating behavior, calorie intake, and nutrient composition ratios. The eating behavior prediction module performs time series analysis based on LSTM networks, detects patterns of rapid change in differential body temperature information, recognizes specific patterns in the cardiovascular response index, and learns and applies user-specific meal patterns. The calorie intake estimation module performs correlation analysis between the magnitude of integral body temperature information and calorie intake, models the relationship between the change patterns of the cardiovascular response index and meal quantity, and performs corrections considering the user's physical characteristics. The nutrient composition ratio analysis module classifies the ratios of protein, carbohydrates, and fats according to the patterns of differential body temperature information, analyzes the temporal change patterns of the cardiovascular response index, and utilizes classification algorithms based on the characteristics of body temperature changes by nutrient. The AI ​​analysis unit enhances prediction accuracy by utilizing various machine learning algorithms and improves the model through continuous learning. Additionally, it may be equipped to perform uncertainty quantification, explainable AI, and anomaly pattern detection.

[0094] The database is responsible for storing multiple user data and maintaining customized models for each user. It stores and manages information such as user profile data, raw biosignal data, processed feature data, customized model parameters, dietary record data, and feedback data. The database provides functions such as data encryption, data backup and recovery, data versioning, and distributed storage.

[0095] The result provision unit is responsible for transmitting the prediction results of the artificial intelligence analysis unit to a user device (200) or linking with a medical institution server (300). The user interface creation module is responsible for visualizing the prediction results, generating customized feedback, and generating notifications. The medical institution linkage module is responsible for converting to a medical standard format, secure communication with the medical institution server (300), and providing detailed analysis data for medical professionals. The result provision unit also includes functions such as support for various platforms, real-time notifications, and periodic reports.

[0096] The edge AI distribution unit is responsible for creating an edge AI unit that includes part or all of the AI ​​analysis unit and transmitting it to a wearable device (210) or a user device (200). The model lightweighting module reduces the model size and increases computational efficiency through model compression, quantization, and knowledge distillation, and transfers knowledge from a complex model to a simple model. The device optimization module performs hardware profiling, resource allocation optimization, and power efficiency optimization to provide optimal performance suitable for the target device. The distribution management module is responsible for version control, remote updates, and performance monitoring. The edge AI unit can be equipped to enable reduced latency, enhanced privacy protection, reduced network dependency, and server load distribution by performing real-time local processing and analysis of biosignals on the wearable device (210) or the user device (200).

[0097] The service server (100) forms a workflow in which a communication unit receives biosignal data from a wearable device (210), a data processing unit preprocesses the received data and extracts features, an artificial intelligence analysis unit predicts meal-related information based on the extracted features, stores raw data, processing results, and prediction results in a database, a result providing unit transmits the prediction results to a user device (200) or links with a medical institution server (300), and an edge artificial intelligence distribution unit distributes an optimized model to the wearable device (210) or the user device (200). The service server (100) has operational characteristics such as scalability, availability, security, and maintainability. The service server (100) of the present invention provides a personalized health management service by analyzing biosignals collected from a wearable device (210) to predict the user's eating behavior, amount of food, and nutrient composition. In particular, by introducing a distributed processing method through an edge artificial intelligence unit, real-time analysis and feedback are possible, and the quality of health management can be improved through linkage with a medical institution. The method for analyzing feeding behavior based on biosignals through the service server (100) is described in detail below.

[0098] The method for analyzing eating behavior based on biosignals through a service server (100) according to an embodiment of the present invention is to predict the eating behavior of a user by analyzing biosignals collected through a wearable device (210) and to analyze the amount of food and nutrient composition, and can provide a personalized health management service by analyzing eating patterns using biosignals such as the user's body temperature, heart rate, and blood pressure.

[0099] The biosignal-based eating behavior analysis method of the present invention relates to a method of operation of a service server (100) that predicts a user's eating behavior, amount of food, and nutrient composition based on biosignal data collected from a wearable device (210), and each step is described in detail below.

[0100] A biosignal-based eating behavior analysis method according to an embodiment of the present invention includes the step of receiving body temperature information, heart rate information, and blood pressure information measured by a body temperature measuring unit, a heart rate measuring unit, and a blood pressure measuring unit equipped in the wearable device (210) from a user device (200) (S110).

[0101] In step (S110), the service server (100) receives body temperature information, heart rate information, and blood pressure information measured by the body temperature measuring unit, heart rate measuring unit, and blood pressure measuring unit equipped in the wearable device (210) from the user device (200).

[0102] The above-mentioned wearable device (210) is worn or attached to the user's body, and the body temperature measuring unit measures the user's body temperature in real time, the heart rate measuring unit measures the user's heart rate, and the blood pressure measuring unit measures the user's blood pressure. These measurements can be performed continuously or at set time intervals.

[0103] The above service server (100) receives biosignal data from a wearable device (210) through a user device (200) (e.g., smartphone, tablet, etc.), and various communication protocols such as Bluetooth, WiFi, and NFC may be utilized in this process.

[0104] In the receiving step (S110), the same type of biosignal may be received from a plurality of wearable devices (210). In this case, the service server (100) may determine whether there is a measurement error by calculating the correlation coefficient of the measurements between the plurality of wearable devices (210). If the correlation coefficient is less than 0.5 to 0.7, it may be determined that there is a measurement error and a self-correction command may be transmitted to the device.

[0105] Next, the method includes the step of generating body temperature derivative information, which is a derivative value on the time axis, and body temperature integral information, which is an integral value relative to the reference body temperature, from the body temperature information (S120).

[0106] In step (S120), the service server (100) generates body temperature derivative information, which is a derivative value over the time axis, and body temperature integral information, which is an integral value relative to the reference body temperature, from the body temperature information.

[0107] The above differential body temperature information is an indicator representing the rate of change of body temperature over time and is utilized for the purpose of detecting changes in body temperature at the start of a meal. The differential body temperature information can be calculated as follows:

[0108]

[0109] Differential body temperature = dT / dt

[0110] Here, T represents body temperature and t represents time.

[0111]

[0112] The above body temperature integral information is an indicator representing the integral value of body temperature relative to a reference body temperature, and is used to measure the cumulative effect of the increase in body temperature caused by a meal. The body temperature integral information can be calculated as follows:

[0113]

[0114] Body temperature integral information = ∫(T-T0)dt

[0115] Here, T represents body temperature, T0 represents reference body temperature, and t represents time.

[0116]

[0117] The pattern of the above differential body temperature information exhibits different characteristics depending on the composition of the consumed nutrients, and in particular, the slope and duration of the differential body temperature information serve as important indicators for classifying the composition ratio of nutrients such as protein, carbohydrates, and fats. For example, if the slope of the differential body temperature information is 0.3°C / 20 min to 0.7°C / 20 min or more and the duration is 15 min to 25 min or more, it can be classified as a high-protein diet.

[0118]

[0119] Next, the method includes the step of generating a cardiovascular response index by calculating the ratio of the changes in the heart rate information and blood pressure information (S130).

[0120] In step (S130), the service server (100) calculates the ratio of changes in heart rate information and blood pressure information to generate a cardiovascular response index.

[0121] The above cardiovascular response index is defined as the ratio of the change in blood pressure to the change in heart rate (△BP / △HR), and is an indicator that quantifies the cardiovascular system's response to a meal. This index is calculated as follows:

[0122]

[0123] Cardiovascular Response Index = △BP / △HR

[0124] Here, △BP represents the change in blood pressure, and △HR represents the change in heart rate.

[0125]

[0126] The above cardiovascular response index is used as an important indicator to determine the amount of food consumed and whether overeating has occurred. In particular, if the cardiovascular response index is less than -0.3, it is determined to be an overeating state, and if it is less than -0.5, it can be determined to be a severe overeating state. In such cases, the service server (100) can generate overeating state information and transmit it to the user device (200).

[0127] The cardiovascular response index can also be used to analyze long-term dietary patterns and to distinguish between breakfast, lunch, dinner, and snacks through patterns of change over time.

[0128]

[0129] Next, the method includes the step of inputting the above-mentioned differential body temperature information, integral body temperature information, and cardiovascular response index into a machine learning model to predict eating behavior, food intake, and / or nutrient composition (S140).

[0130] In step (S140), the service server (100) inputs the differential body temperature information, the integral body temperature information, and the cardiovascular response index into a machine learning model to predict eating behavior, food quantity, and / or nutrient composition.

[0131] The above machine learning model is constructed by learning biometric information collected from multiple users and actual eating behavior data, and the service server (100) periodically updates it. These updates are carried out in a way that improves the accuracy of the model through comparison between actual data and predicted values.

[0132] The above machine learning model can be built to be personalized to reflect the characteristics of each user. The service server (100) can generate a user profile for a specific user selected among a plurality of users, and can adjust the model weights to be personalized by comparing the actual eating behavior data and predicted values ​​for the first 5 to 10 days.

[0133] The aforementioned machine learning model can estimate calorie intake based on the magnitude of body temperature integral information and generate nutrient composition ratio information by classifying the ratios of protein, carbohydrates, and fats according to the patterns of body temperature differential information. This analysis helps improve the user's dietary habits by providing information on the qualitative aspects of the meal.

[0134]

[0135] Next, the method includes the step of transmitting the above prediction result to a user device (200) or linking it with medical data from a medical institution (S150).

[0136] In step (S150), the service server (100) transmits the prediction result to the user device (200) or links with the medical data of the medical institution.

[0137]

[0138] The above prediction results include eating behavior, food intake, nutrient composition ratio, etc., and can be visually provided to the user through an application of the user device (200). This information can be used to improve the user's eating habits and manage their health.

[0139] Linkage with the medical data of the aforementioned medical institution is performed with the user's consent, thereby enabling medical professionals to comprehensively analyze the user's dietary habits and related health information. Additionally, the service server (100) can share anonymized biometric information and meal pattern data with a medical research institution with the user's consent, and update the machine learning model by receiving an improved analysis algorithm from the institution.

[0140] The above result provision step may additionally include a step of distinguishing breakfast, lunch, dinner, and snacks by analyzing the change patterns of body temperature information, heart rate information, and blood pressure information over time, and a step of generating dietary habit analysis information based thereon and transmitting it to a user device (200).

[0141] The present invention includes an edge artificial intelligence unit that performs local processing not only on a cloud-based service server (100) but also on a wearable device (210) or a user device (200).

[0142] The service server (100) can generate an edge artificial intelligence unit that includes part or all of a machine learning model and transmit it to a user device (200) or a wearable device (210). The edge artificial intelligence unit can process and analyze biosignals collected from the wearable device (210) locally in real time to provide immediate feedback without an internet connection.

[0143] The edge artificial intelligence unit is implemented to be optimized for the limited computing resources of the wearable device (210) and is optimized to the hardware performance of the wearable device (210) through quantization and model compression techniques. This enables efficient artificial intelligence computation even on the wearable device (210).

[0144] The service server (100) receives the analysis results processed through the edge artificial intelligence unit from the wearable device (210), performs cloud-based advanced analysis based on the results, and then creates an updated edge artificial intelligence unit and retransmits it to the wearable device (210). This distributed processing method enables the use of both real-time processing and advanced analysis.

[0145] By generating a user-customized edge artificial intelligence unit based on user profiles and corrected model weights and transmitting it to a user device (200) or wearable device (210), personalized eating behavior prediction and analysis can be performed locally.

[0146] The biosignal-based eating behavior analysis method according to an embodiment of the present invention can provide a personalized health management service by analyzing biosignals collected from a wearable device (210) to predict the user's eating behavior, amount of food, and nutrient composition. In particular, it has a technical feature of analyzing eating patterns using differential body temperature information, integral body temperature information, and cardiovascular response index calculated from biosignals such as body temperature, heart rate, and blood pressure.

[0147] In addition, by introducing a distributed processing method through the edge artificial intelligence unit, real-time analysis and feedback are possible even without an internet connection, and the quality of health management can be improved through linkage with medical institutions. Through this, it can contribute not only to personal health management but also to the field of medical research, and the utility value of the wearable device (210) can be enhanced.

[0148]

[0149] According to another embodiment of the present invention, a personalized health management service can be provided by analyzing the eating pattern using biosignals collected through a wearable device (210) to predict the user's eating behavior and to analyze the amount of food and nutrient composition.

[0150] A biosignal-based feeding behavior analysis method according to another embodiment of the present invention relates to a method in which a wearable device (210) collects bio-information and an edge artificial intelligence unit analyzes it to predict feeding behavior, and each step is described in detail below.

[0151] First, the wearable device (210) includes the step of collecting body temperature information, heart rate information, and blood pressure information, respectively, through a body temperature measuring unit, a heart rate measuring unit, and a blood pressure measuring unit (S210).

[0152] In step (S210), the wearable device (210) collects body temperature information, heart rate information, and blood pressure information, respectively, through a body temperature measuring unit, a heart rate measuring unit, and a blood pressure measuring unit.

[0153] The above-mentioned wearable device (210) is worn or attached to the user's body in the form of a finger-worn device (210a), a wrist-worn device (210b), an ear-worn device (210c), a chest-worn device (210d), a glasses-worn device (210e), etc. to measure biosignals. Such measurements can be performed continuously or at set time intervals.

[0154] The above-described wearable device (210) can determine whether there is a measurement error by calculating the correlation coefficient of the measurement values ​​between multiple devices. If the correlation coefficient is less than 0.5 to 0.7, it can determine that there is a measurement error and perform self-correction.

[0155] In addition, by utilizing exercise information collected through an accelerometer equipped in a wearable device (210), the edge artificial intelligence unit can perform a preprocessing process to distinguish between exercise-induced body temperature rise and meal-induced body temperature rise.

[0156]

[0157] Next, the edge artificial intelligence unit of the wearable device (210) includes the step of generating body temperature derivative information, which is a derivative value on the time axis, and body temperature integral information, which is an integral value relative to the reference body temperature, from the body temperature information (S220).

[0158] In step (S220), the edge artificial intelligence unit of the wearable device (210) generates body temperature derivative information, which is a derivative value on the time axis, and body temperature integral information, which is an integral value relative to the reference body temperature, from the body temperature information.

[0159] The above differential body temperature information is an indicator representing the rate of change of body temperature over time and is utilized for the purpose of detecting changes in body temperature at the start of a meal. The differential body temperature information can be calculated as follows:

[0160]

[0161] Differential body temperature = dT / dt

[0162] Here, T represents body temperature and t represents time.

[0163]

[0164] The above body temperature integral information is an indicator representing the integral value of body temperature relative to a reference body temperature, and is used to measure the cumulative effect of the increase in body temperature caused by a meal. The body temperature integral information can be calculated as follows:

[0165]

[0166] Body temperature integral information = ∫(T-T0)dt

[0167] Here, T represents body temperature, T0 represents reference body temperature, and t represents time.

[0168]

[0169] The above body temperature integration information is a cumulative value of the increase in body temperature due to heat generation inducing digestion, wherein a wearable device (210) continuously measures the user's body temperature, and an edge artificial intelligence unit calculates the rate of change in body temperature over time from the measured body temperature data. If the calculated rate of change in body temperature exceeds a predetermined threshold, it can be determined as rapid food intake.

[0170]

[0171] In particular, if a period of 10 minutes or more during which the rate of increase in body temperature is 0.3℃ / 10 minutes or more is confirmed within a measurement period of 30 to 40 minutes, fast eating status information can be generated and output.

[0172]

[0173] Next, the edge artificial intelligence unit calculates the ratio of the change in heart rate information and blood pressure information to generate a cardiovascular response index (S230).

[0174] In step (S230), the edge artificial intelligence unit of the wearable device (210) calculates the ratio of changes in heart rate information and blood pressure information to generate a cardiovascular response index.

[0175] The above cardiovascular response index is defined as the ratio of the change in blood pressure to the change in heart rate (△BP / △HR), and is an indicator that quantifies the cardiovascular system's response to a meal. This index is calculated as follows:

[0176]

[0177] Cardiovascular Response Index = △BP / △HR

[0178] Here, △BP represents the change in blood pressure, and △HR represents the change in heart rate.

[0179]

[0180] The above cardiovascular response index is used as an important indicator to determine the amount of food consumed and whether overeating has occurred. In particular, if the cardiovascular response index is less than -0.3, it is determined to be an overeating state, and overeating state information is generated and output; if it is less than -0.5, it can be determined to be a severe overeating state.

[0181]

[0182] Next, the edge artificial intelligence unit inputs the body temperature differential information, body temperature integral information, and cardiovascular response index into a machine learning model to predict eating behavior, food intake, and / or nutrient composition (S240).

[0183] In step (S240), the edge artificial intelligence unit of the wearable device (210) inputs differential body temperature information, integral body temperature information, and cardiovascular response index into a machine learning model to predict eating behavior, food quantity, and / or nutrient composition.

[0184] The above machine learning model is equipped with a hybrid structure of Long Short-Term Memory (LSTM) and Temporal Convolutional Network, and is configured to simultaneously analyze the standard deviation of differential body temperature information, the coefficient of variation of cardiovascular response index, and the slope of integral body temperature information within a time window of 10 to 20 minutes.

[0185] The above edge artificial intelligence unit estimates calorie intake based on the magnitude of body temperature integral information and classifies the ratios of protein, carbohydrates, and fat based on the pattern of body temperature differential information. In particular, if the gradient of the body temperature differential information is 0.3°C / 20 min to 0.7°C / 20 min or more and the duration is 15 min to 25 min or more, it is classified as a high-protein diet to generate high-protein diet information, and if the gradient of the body temperature differential information is 0.1°C / 20 min to 0.3°C / 20 min or less and the duration is 10 min to less than 15 min, it is classified as a high-fat diet to generate high-fat diet information.

[0186]

[0187] Next, the wearable device (210) provides the prediction result to a user through an output unit or transmits it to a user device (200) or a service server (100) (S250).

[0188] In step (S250), the wearable device (210) provides the prediction result to the user through the output unit or transmits it to the user device (200) or service server (100).

[0189] The above prediction results include eating behavior, amount of food, nutrient composition ratio, etc., and may be provided directly to the user through the output of the wearable device (210) or transmitted to the user device (200) or service server (100).

[0190] The wearable device (210) transmits the eating behavior prediction results, meal amount estimation results, and nutrient composition ratio analysis results generated by the edge artificial intelligence unit, along with body temperature information, heart rate information, and blood pressure information measured at regular intervals in a 24-hour cycle, to the service server (100), and can automatically update by receiving an updated edge artificial intelligence unit with improved analysis accuracy from the service server (100).

[0191] Meanwhile, the wearable device (210) may receive and install an edge artificial intelligence unit from the service server (100) or the user device (200). The received edge artificial intelligence unit is configured to include a lightweight artificial intelligence model optimized for the hardware performance of the wearable device (210).

[0192] This edge artificial intelligence unit can process and analyze biosignals in real time from the wearable device (210) and provide immediate feedback without an internet connection. The edge artificial intelligence unit is implemented to be optimized for the limited computing resources of the wearable device (210) and is optimized to the hardware performance of the wearable device (210) through quantization and model compression techniques.

[0193] A biosignal-based eating behavior analysis method according to an embodiment of the present invention can provide a personalized health management service by analyzing biosignals collected from a wearable device (210) to predict the user's eating behavior, food intake, and nutrient composition. In particular, it is configured to analyze eating patterns using differential body temperature information, integral body temperature information, and cardiovascular response index calculated from biosignals such as body temperature, heart rate, and blood pressure.

[0194]

[0195] Preferred embodiments of the present invention are disclosed in this specification and drawings. Although specific terms have been used, they are used only in a general sense to facilitate the explanation of the technical content of the invention and to aid in understanding the invention, and are not intended to limit the scope of the invention. It is obvious to those skilled in the art that, in addition to the embodiments disclosed herein, other variations based on the technical concept of the present invention are possible.

Claims

1. A method of operation of a service server that receives biosignal data from at least one wearable device worn or attached to a user's body and predicts eating behavior, wherein The above service server, A step of receiving body temperature information, heart rate information, and blood pressure information measured by a body temperature measuring unit, a heart rate measuring unit, and a blood pressure measuring unit equipped in the wearable device from a user device; A step of generating body temperature derivative information, which is a derivative value on the time axis, and body temperature integral information, which is an integral value relative to the reference body temperature, from the above body temperature information; A step of generating a cardiovascular response index by calculating the ratio of the change amounts of the heart rate information and blood pressure information above; A step of inputting the above-mentioned differential body temperature information, integral body temperature information, and cardiovascular response index into a machine learning model to predict eating behavior, food intake, and / or nutrient composition; and A biosignal-based eating behavior analysis method characterized by including the step of transmitting the above prediction result to a user device or linking it with medical data from a medical institution.

2. In Paragraph 1, The above service server, A step of generating an edge artificial intelligence unit comprising part or all of the machine learning model; and The method further includes the step of transmitting the generated edge artificial intelligence unit to the user device or the wearable device; A method for analyzing eating behavior based on biosignals, characterized in that the edge artificial intelligence unit is configured to process and analyze biosignals collected from the wearable device locally in real time to provide immediate feedback without an internet connection.

3. In Paragraph 2, A method for analyzing feeding behavior based on biosignals, characterized in that the edge artificial intelligence unit transmitted by the service server is configured to be optimized for the limited computing resources of a wearable device and is optimized to match the hardware performance of the wearable device through quantization and model compression techniques.

4. In Paragraph 2, A method for analyzing eating behavior based on biosignals, further comprising the step of: the service server receiving an analysis result processed through an edge artificial intelligence unit from the wearable device, performing a cloud-based advanced analysis based thereon, generating an updated edge artificial intelligence unit, and retransmitting it to the wearable device.

5. In Paragraph 1, A method for analyzing eating behavior based on biosignals, characterized by further including the step of the service server learning biometric information collected from multiple users and actual eating behavior data to periodically update the machine learning model.

6. In Paragraph 5, The above service server, A step of constructing a personalized prediction model by generating a user-specific profile of a specific user selected among the plurality of users above; and A biosignal-based eating behavior analysis method characterized by further including the step of comparing actual eating behavior data and predicted values ​​during the initial 5 to 10 days to adjust model weights in a personalized manner.

7. In Paragraph 6, A biosignal-based eating behavior analysis method further comprising the step of: the service server generating a user-customized edge artificial intelligence unit based on the user-specific profile and corrected model weights and transmitting it to a user device or wearable device so that personalized eating behavior prediction and analysis can be performed locally.

8. In Paragraph 1, A biosignal-based eating behavior analysis method characterized by further including the step of defining the above cardiovascular response index as the ratio of the change in blood pressure to the change in heart rate (△BP / △HR), and the service server determining an overeating state when the above cardiovascular response index is less than -0.3, generating overeating state information, and transmitting it to a user device.

9. In Paragraph 8, A biosignal-based eating behavior analysis method characterized by the above service server being configured to determine an overeating state when the cardiovascular response index is less than -0.5, generate overeating state information, and transmit it to the user device.

10. In Paragraph 1, The above service server, A step of estimating calorie intake according to the magnitude of the above-mentioned body temperature integral information, and generating nutrient composition ratio information by classifying the ratios of protein, carbohydrates, and fats according to the pattern of the above-mentioned body temperature differential information; and A biosignal-based eating behavior analysis method characterized by further including the step of transmitting the estimated calorie amount and nutrient composition ratio information to the user device.

11. In Paragraph 10, A method for analyzing eating behavior based on biosignals, further comprising the step of: the service server generating an edge artificial intelligence unit including a lightweight artificial intelligence model for calorie estimation and nutrient composition ratio classification and transmitting it to the wearable device, thereby enabling the wearable device to analyze intake calories and nutrient composition ratios in real time without an internet connection and provide immediate feedback.

12. In Paragraph 10, The step of generating the above-mentioned nutrient composition ratio information is, A biosignal-based eating behavior analysis method characterized by being configured to classify the above-mentioned differential body temperature information as a high-protein diet and generate and transmit the high-protein diet information to a user device when the gradient of the above-mentioned differential body temperature information is 0.3°C / 20 min to 0.7°C / 20 min or more and the duration is 15 min to 25 min or more, or to classify the above-mentioned differential body temperature information as a high-fat diet and generate and transmit the high-fat diet information to a user device when the gradient of the above-mentioned differential body temperature information is 0.1°C / 20 min to 0.3°C / 20 min or less and the duration is 10 min to less than 15 min.

13. In Paragraph 1, A method for analyzing eating behavior based on biosignals, further comprising the step of: the service server determining that there is a measurement error when the correlation coefficient between measurements of multiple wearable devices is less than 0.5 to 0.7, transmitting a self-correction command to the device, and receiving the self-correction result to utilize it for improving device performance.

14. In Paragraph 13, A method for analyzing eating behavior based on biosignals, further comprising the step of: the service server generating a distributed edge artificial intelligence unit for self-correction and data synchronization among the plurality of wearable devices and transmitting it to each wearable device, thereby enabling the plurality of wearable devices to communicate with each other to autonomously correct and synchronize measurement values.

15. In Paragraph 1, The above service server, A step of distinguishing breakfast, lunch, dinner, and snacks by analyzing the change patterns over time of the above body temperature information, heart rate information, and blood pressure information; and A biosignal-based eating behavior analysis method characterized by further including the step of generating dietary habit analysis information based on the above classification results and transmitting it to a user device.

16. In Paragraph 1, The above service server, A step of sharing anonymized biometric information and meal pattern data with a medical research institution with the user's consent; and A method for analyzing feeding behavior based on biosignals, further comprising the step of receiving an improved analysis algorithm from the medical research institution and updating the machine learning model.

17. In Paragraph 16, A method for analyzing eating behavior based on biosignals, further comprising the step of: generating an updated edge artificial intelligence unit based on an improved analysis algorithm received from the medical research institution by the service server and transmitting it to the wearable device so that an analysis reflecting the latest medical research results can be performed locally.

18. A service server that analyzes biosignals and predicts eating behavior by communicating with at least one wearable device worn or attached to a user's body and / or a user device connected to said wearable device, A communication unit that receives body temperature information, heart rate information, and blood pressure information from the above-mentioned wearable device; A data processing unit that generates differential body temperature information and integral body temperature information from the above body temperature information, and generates a cardiovascular response index from the above heart rate information and blood pressure information; An artificial intelligence analysis unit that receives the above-mentioned differential body temperature information, integral body temperature information, and cardiovascular response index as input to predict eating behavior, calorie intake, and nutrient composition ratios; A database that stores multiple user data and maintains a customized model for each user; and A service server characterized by including a result providing unit that transmits prediction results to a user device or links with a medical institution server.

19. In Paragraph 18, The service server further comprises an edge artificial intelligence distribution unit that generates an edge artificial intelligence unit including part or all of the artificial intelligence analysis unit and transmits it to the wearable device or the user device, wherein the edge artificial intelligence unit is configured to perform real-time local processing and analysis of biosignals by being optimized for the hardware performance of the wearable device or the user device.

20. A method in which a wearable device worn or attached to a user's body collects biometric information and an edge artificial intelligence unit included in the wearable device analyzes the collected biometric information to predict eating behavior, The above-mentioned wearable device, A step of collecting body temperature information, heart rate information, and blood pressure information, respectively, through a body temperature measuring unit, a heart rate measuring unit, and a blood pressure measuring unit; A step in which the edge artificial intelligence unit of the above-described wearable device generates body temperature derivative information, which is a derivative value on the time axis, and body temperature integral information, which is an integral value relative to a reference body temperature, from the body temperature information; A step in which the edge artificial intelligence unit calculates the ratio of the change amounts of the heart rate information and blood pressure information to generate a cardiovascular response index; A step in which the edge artificial intelligence unit inputs the body temperature differential information, body temperature integral information, and cardiovascular response index into a machine learning model to predict eating behavior, food quantity, and / or nutrient composition; and A method for analyzing eating behavior based on biosignals, characterized by including the step of the wearable device providing the prediction result to a user through an output unit or transmitting it to a user device or a service server.

21. In Paragraph 20, A method for analyzing eating behavior based on biosignals, characterized in that the above-mentioned wearable device is a finger-worn device, a wrist-worn device, an ear-worn device, a chest-worn device, or a glasses-worn device, and further comprises the step of performing self-correction when the correlation coefficient of the measured values ​​between each device is less than 0.5 to 0.

7.

22. In Paragraph 20, The above body temperature integration information is the cumulative value of the increase in body temperature due to thermogenesis induced by digestion, A step in which the above-mentioned wearable device continuously measures the user's body temperature; A step in which the edge artificial intelligence unit of the wearable device calculates the rate of change in body temperature over time from the measured body temperature data; and A biosignal-based eating behavior analysis method characterized by being generated through the step in which the edge artificial intelligence unit of the wearable device determines rapid food intake when the calculated rate of change in body temperature exceeds a predetermined threshold.

23. In Paragraph 22, A method for analyzing eating behavior based on biosignals, further comprising the step of generating and outputting fast eating state information when the edge artificial intelligence unit of the wearable device confirms that the body temperature rise rate is 0.3℃ / 10 minutes or more and lasts for 10 minutes or more within a measurement period of 30 to 40 minutes.

24. In Paragraph 20, A biosignal-based eating behavior analysis method characterized by further including the step of: the above cardiovascular response index is defined as the ratio of the change in blood pressure to the change in heart rate (△BP / △HR), and the edge artificial intelligence unit of the wearable device determines that the cardiovascular response index is in an overeating state when it is less than -0.3, and generates and outputs overeating state information.

25. In Paragraph 24, A biosignal-based eating behavior analysis method characterized in that the edge artificial intelligence unit of the above-described wearable device is configured to determine an overeating state when the cardiovascular response index is less than -0.5, and to generate and output overeating state information.

26. In Paragraph 20, A method for analyzing eating behavior based on biosignals, characterized by further including a preprocessing step in which an edge artificial intelligence unit of the wearable device distinguishes between an exercise-induced body temperature rise and a meal-induced body temperature rise by utilizing exercise information collected through an accelerometer unit equipped in the wearable device.

27. In Paragraph 20, A biosignal-based feeding behavior analysis method characterized by the above machine learning model being equipped with a hybrid structure of Long Short-Term Memory LSTM and Temporal Convolutional Network, and being configured to simultaneously analyze the standard deviation of differential body temperature information, the coefficient of variation of cardiovascular response index, and the gradient of integral body temperature information within a time window of 10 to 20 minutes.

28. In Paragraph 20, A method for analyzing eating behavior based on biosignals, characterized by further including the step of the edge artificial intelligence unit of the wearable device estimating calorie intake according to the magnitude of the body temperature integral information and classifying the ratios of protein, carbohydrates, and fat according to the pattern of the body temperature differential information.

29. In Paragraph 28, A method for analyzing eating behavior based on biosignals, characterized in that the edge artificial intelligence unit of the wearable device is configured to classify the differential body temperature information as a high-protein diet and generate high-protein diet information and transmit the high-protein diet information to the user device or service server when the gradient of the differential body temperature information is 0.3°C / 20 minutes to 0.7°C / 20 minutes or more and the duration is 15 minutes to 25 minutes or more, and classify the differential body temperature information as a high-fat diet and generate high-fat diet information and transmit the high-fat diet information to the user device or service server when the gradient of the differential body temperature information is 0.1°C / 20 minutes to 0.3°C / 20 minutes or less and the duration is 10 minutes to less than 15 minutes.

30. In Paragraph 20, A method for analyzing eating behavior based on biosignals, further comprising the step of receiving and installing an edge artificial intelligence unit from a service server or a user device on the wearable device, wherein the received edge artificial intelligence unit is configured to include a lightweight artificial intelligence model optimized for the hardware performance of the wearable device.

31. In Paragraph 20, A method for analyzing eating behavior based on biosignals, further comprising the step of: the wearable device transmitting body temperature information, heart rate information, and blood pressure information measured at regular intervals in a 24-hour cycle to the service server, along with the eating behavior prediction results, meal amount estimation results, and nutrient composition ratio analysis results generated by the edge artificial intelligence unit; and receiving an updated edge artificial intelligence unit with improved analysis accuracy from the service server and automatically updating it.

Images