Omnibearing personalized nutrition management method and system, intelligent scale and server

Abstract

The invention relates to the field of health management, and discloses a comprehensive personalized nutrition management method, which comprises the following steps: acquiring multi-dimensional health data of a user through user side equipment, including gene information, living habits, exercise frequency and food intake information; analyzing the multi-dimensional health data by using an AI-driven health assessment model, and generating an assessment result of the health state of the user and potential risk factors; and generating a personalized diet plan by using a machine learning algorithm according to the health assessment result. Through multi-dimensional data acquisition and dynamic analysis, the health state of the user is comprehensively monitored, the diet plan is adjusted in real time, and the problem of insufficient static property in the prior art is solved; gene information and living habits are integrated, and a highly personalized effect is achieved; the diet recommendation accuracy is improved through data closed-loop optimization, and the defect that an existing scheme is insufficient in optimization capability is overcome; tLS encryption and distributed storage are adopted, data security is guaranteed, and user trust is enhanced.

Description

Technical Field

[0001] The present invention relates to the field of health management, and specifically to an all-round personalized nutrition management method, system, smart scale and server. Background Art

[0002] Existing health management systems mostly rely on questionnaires or manual input to obtain basic user data, such as height, weight, and eating habits. This method cannot dynamically reflect changes in the user's health status, such as weight fluctuations or dietary intake, and at the same time, it is also easy to affect the accuracy of the evaluation results due to data omission or input errors.

[0003] Most systems generate dietary plans through simple rule matching, lacking in-depth analysis of multi-dimensional data such as the user's genetic information and metabolic characteristics. This general-purpose recommendation method is difficult to meet the personalized needs of chronic disease patients or special populations, resulting in poor actual effects of the plan.

[0004] Existing technologies also have deficiencies in dynamic monitoring and cannot track the user's dietary behavior in real time. For example, high-calorie or high-fat foods ingested by the user are often not captured by the system, resulting in a disconnect between dietary guidance and actual behavior and weak adjustment capabilities.

[0005] In addition, user feedback data is not effectively utilized, the system lacks a closed-loop optimization mechanism, and it is difficult to adjust the recommendation plan according to changes in the user's health needs. The centralized storage method also provides insufficient privacy protection for user data during transmission and storage, posing a risk of leakage. Summary of the Invention

[0006] In view of the deficiencies of the existing technologies, the present invention provides an all-round personalized nutrition management method, system, smart scale and server, which solves the problems of insufficient personalization, limited real-time monitoring ability, and lack of guarantee for data security in existing health management systems.

[0007] To achieve the above objectives, the present invention is realized through the following technical solutions: An all-round personalized nutrition management method, including the following steps: S1. Obtain multi-dimensional health data of the user through the user terminal device, including genetic information, living habits, exercise frequency, and food intake information; S2. Analyze the above multi-dimensional health data using an AI-driven health assessment model to generate an evaluation result of the user's health status and potential risk factors; S3. Generate a personalized dietary plan using a machine learning algorithm based on the health assessment result; S4. Collect the user's actual food intake data and cooking methods through the smart scale and upload them to the server in real time; S5. Dynamically optimize the personalized dietary plan based on user feedback information and real-time collected data.

[0008] Preferably, the health assessment model in step S2 comprehensively adopts a genetic algorithm and a deep learning algorithm, and obtains personalized health risk analysis results based on big data training.

[0009] Preferably, in step S5, by regularly collecting the user's usage feedback and changes in health indicators, a reinforcement learning mechanism is used to dynamically adjust the dietary recommendation strategy.

[0010] Preferably, an all-round personalized nutrition management system includes: A user-side device for collecting multi-dimensional health data of the user; An intelligent scale for collecting the user's food intake data and cooking methods; A cloud server, including: a health assessment module for generating a health status assessment result based on the user's health data; a dietary plan generation module for generating a personalized dietary plan according to the health assessment result; a data optimization module for optimizing the dietary plan based on the user's feedback and real-time collected data; A client application for interacting with the user and displaying the dietary plan and health assessment result.

[0011] Preferably, the cloud server further includes a virtual assistant module, which combines natural language processing technology and an emotion analysis engine to push personalized health suggestions and recipe recommendations according to the user's preferences and emotional state.

[0012] Preferably, the user-side device includes a mobile application for collecting the user's health data and displaying the personalized dietary plan and health status assessment result.

[0013] An intelligent scale includes: A food weight sensor for measuring the weight of the user's food ingredients; A cooking mode recognition module for identifying the user's cooking method; A data transmission module for uploading the food weight and cooking mode information to the cloud server in real time; A built-in processing module for formatting the collected data and binding it to the user ID.

[0014] Preferably, the data transmission module encrypts the data using an encryption protocol to ensure the security of the food intake data during transmission.

[0015] A cloud server for all-round personalized nutrition management includes: A health assessment module for generating a health status assessment result based on the user's multi-dimensional health data; A dietary plan generation module for generating a personalized dietary plan according to the health assessment result; A data optimization module for optimizing the dietary plan based on user feedback and real-time collected data; A data storage module for storing user health data and dietary plans and setting multi-level permission control.

[0016] Preferably, a distributed storage architecture is adopted to encrypt and store user health data, and access permissions are controlled through a multi-layer authentication mechanism.

[0017] The present invention provides an all-round personalized nutrition management method, system, smart scale and server. It has the following beneficial effects: 1. The present invention adopts a technical solution based on multi-dimensional data collection and dynamic analysis, achieving the technical effect of comprehensively monitoring the user's health status and adjusting the dietary plan in real time. Compared with the technical solutions in the prior art that rely on static questionnaires and fixed dietary templates, it solves the deficiency of being difficult to dynamically reflect the user's health changes and actual eating behaviors.

[0018] 2. By integrating the user's genetic information, living habits and real-time diet data, the present invention can accurately generate a dietary plan that meets personal needs, achieving the technical effect of being tailored to each individual and providing precise guidance. Compared with the prior art health management solutions with strong generality and lack of pertinence, the present invention solves the technical deficiency that the individual differences of users cannot be fully considered.

[0019] 3. The present invention uses smart hardware to collect the user's actual diet data in real time, and combines reinforcement learning technology to continuously optimize dietary recommendations, forming a closed-loop management system for data collection, analysis and feedback, achieving the technical effect of continuously improving the accuracy of the dietary plan. Compared with the problem in the prior art that it is difficult to adjust the recommendation scheme based on user feedback and behavior, the present invention overcomes the defect of insufficient system optimization ability.

[0020] 4. The present invention adopts a technical solution of TLS protocol encryption and distributed storage in data transmission and storage, achieving the technical effect of protecting user privacy and sensitive data. Compared with the problems of data leakage risk and weak privacy protection mechanism in the prior art, the present invention effectively solves the technical shortcoming of insufficient user data security and significantly enhances user trust. Brief Description of the Drawings

[0021] Figure 1 It is a schematic diagram of the step flow of the method in the present invention. Detailed Embodiments

[0022] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the specification of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0023] Embodiment: The method of the present invention collects multi-dimensional health data of users, including genetic information, living habits, dietary behaviors, etc., uses an AI-driven health assessment model to analyze the health status of users and identify potential risks, generates personalized dietary plans according to the assessment results, and realizes real-time diet monitoring and dynamic optimization through intelligent devices. This method continuously optimizes the accuracy and personalization of health advice through continuous user feedback and reinforcement learning technology, providing users with scientific and real-time health management guidance. The specific steps are as follows: User data collection Basic data collection: Users submit basic health data through the client application, including but not limited to parameters such as age, gender, height, and weight. The client interface provides form-based input and user-friendly guided input functions.

[0024] Genetic data collection Genetic data is obtained through an external genetic testing service. Users upload genetic testing report files (such as TXT, CSV formats), and the system uses a built-in parsing module to extract gene-related feature data (such as metabolic capacity, food sensitivity, disease susceptibility, etc.).

[0025] Collection of living habits and health records The client collects data on users' daily living habits through a questionnaire module, including behavioral data such as dietary preferences, work and rest times, alcohol and smoking situations, and exercise frequencies. This module supports linkage with wearable devices (such as sports bracelets) to automatically synchronize exercise data.

[0026] The implementation of this step is directly related to the quality of the input data of the subsequent health assessment model. The comprehensiveness and accuracy of user data are of great significance for generating a scientific health assessment report and a customized dietary plan. The user data collection involved in this step not only includes basic health data but also extends to multi-dimensional contents such as genetic information, living habits, and dietary behaviors. Through the coordinated cooperation of various data collection methods and devices, accurate records of users' health conditions are achieved, providing a reliable data basis for the subsequent steps.

[0027] In this embodiment, user data collection involves multiple aspects, including basic data collection, gene information collection, and collection of lifestyle and health records. Among them, the basic data mainly serves as the basic variables for subsequent analysis, while the gene information and lifestyle data are input as key parameters for generating personalized dietary plans.

[0028] Basic data collection: Generally, users enter their personal basic health data through the client application. Specifically, these data include but are not limited to the following: Basic physical signs information such as the user's age, gender, height, and weight; Current physical condition data, such as body fat percentage (measured by a smart device) or basal metabolic rate (which can be estimated by a formula).

[0029] In a possible implementation, the basal metabolic rate can be calculated by the following formula: BMR = 10×W + 6.25×H - 5×A + S Where, W represents the user's weight (kg), H represents the user's height (cm), A represents the user's age (years), S is the gender factor, with a value of +5 for males and -161 for females.

[0030] The above data is collected through the form-based interface of the client. To improve the accuracy of user input, the system designs various data verification mechanisms. For example, the input range of the user's height and weight is set to a common reasonable interval (for example, height between 120 cm and 250 cm, weight between 30 kg and 200 kg). When the input exceeds the range, the system will pop up a prompt.

[0031] Gene information collection: In some embodiments, the present invention introduces gene information as an important part of data collection. Gene data can be parsed through the gene test reports uploaded by users (such as in TXT or CSV format). These reports usually come from third-party gene testing services, such as the gene polymorphism results obtained by users after testing.

[0032] As an option, gene information is mainly used to evaluate the user's metabolic ability, food sensitivity (such as lactose intolerance), and susceptibility to chronic diseases. Specifically, the system extracts key features by parsing the SNP data (Single Nucleotide Polymorphism) of gene polymorphism sites. For example, the SNP site rs4988235 is related to lactose intolerance, and the system determines the user's lactose metabolism ability by detecting whether the user's genotype is C / C.

[0033] In a possible implementation, the system has a built-in parsing module that can automatically recognize the formats of common gene detection reports and extract the following key gene locus information: Gene loci related to metabolic capacity (such as rs9939609 of the FTO gene); Gene loci related to disease susceptibility (such as rs7903146 of the TCF7L2 gene, used to evaluate the risk of diabetes).

[0034] These gene feature data will be standardized and stored in the user's health record on the server side, providing input for the subsequent health assessment model.

[0035] Collection of lifestyle and health records: In this embodiment, the data collection of lifestyle and health records is completed through the questionnaire module in the client. The user needs to answer multiple questions about lifestyle, including: Dietary habits, such as favorite food types, dietary preferences (such as salty or sweet), and whether there are any dietary taboos (such as allergens).

[0036] Daily routine, such as sleep time (the specific time range can be selected for input) and sleep quality (rated by a scale).

[0037] Exercise habits, such as the weekly exercise frequency, duration, and intensity (such as light, moderate, or heavy).

[0038] As a possible implementation, the client can be integrated with external wearable devices (such as smart bracelets) to automatically synchronize the user's exercise data. For example, the system can calculate the user's exercise amount based on the step count and heart rate records of the bracelet, and calculate the energy consumed during the activity in combination with the MET (metabolic equivalent of task) formula: E = MET × W × T Where: E represents the total energy consumed during the activity (kcal); MET represents the metabolic equivalent of task for the activity (for example, 3.5 for walking); W represents the body weight (kg); T represents the activity time (hours).

[0039] In some embodiments, to ensure the authenticity of the user's answers, the system designs a dynamic questionnaire adjustment mechanism. For example, if the user reports "exercising 30 minutes every day" but no relevant exercise data is detected through the records of the wearable device, the system will prompt the user to verify.

[0040] Data upload and storage: After data collection is completed, all user data will be uploaded to the cloud server through the encrypted communication module. As an option, the encryption protocol can adopt TLS1.3 to ensure the security of data during transmission. Meanwhile, the system normalizes the data on the server side and assigns a unique user ID to achieve efficient management and subsequent association of the data.

[0041] Through the above steps, the present invention has successfully achieved the comprehensive collection of multi-dimensional health data of users. The collected data will be used as the direct input of the health assessment model, ensuring the accuracy and scientific nature of the assessment results and laying a data foundation for the generation of personalized dietary plans.

[0042] The health assessment system runs an AI-driven health assessment model on the server side. The model mainly includes the following modules: Health status assessment module: Using the multi-dimensional data input of users, through the combined analysis of a deep neural network (DNN) and a logistic regression model, it assesses the current health status (such as BMI status, cardiovascular risk level, fat metabolism ability).

[0043] Risk prediction module: Based on genetic data and health history records, it applies genetic algorithms to analyze possible health risks, such as diabetes, obesity, or chronic disease risks. Model input: User health data; Model output: Health assessment report (including current health status, risk warnings).

[0044] In the above steps, multi-dimensional health data of users have been collected, and these data will be used as the basic input for health assessment in subsequent steps. The core of health assessment lies in obtaining the current health status and potential health risks of users through multi-level and multi-angle analysis of user health data, providing a reliable basis for the formulation of subsequent personalized dietary plans. Health assessment not only involves data calculation and analysis, but also needs to combine the user's genetic information, basic data, and lifestyle data, and use a series of algorithms and models for comprehensive processing. The implementation of this step is a key link of the present invention, and its implementation method directly affects the subsequent dietary recommendation quality and user satisfaction.

[0045] In this embodiment, the process of health assessment depends on the AI-driven health assessment model on the cloud server side. This model comprehensively uses deep learning algorithms, logistic regression analysis, and genetic algorithms to comprehensively process and analyze the data uploaded by users.

[0046] Generally, the health assessment model will receive multiple data dimensions collected in the above steps. Including but not limited to: The user's basic health data, such as height, weight, gender, age, etc.; Genetic data, such as gene locus characteristics related to metabolism; Lifestyle data, including diet preferences, exercise intensity, etc.

[0047] In a possible implementation, the system first normalizes these input data to eliminate the dimensional differences between different data dimensions. The normalization formula can adopt the following expression: where x represents the original data value, x min and x max are respectively the minimum and maximum values of this data dimension, x ′ is the result after normalization.

[0048] Through this processing, the values of all input data are normalized to the interval [0, 1], which is convenient for the subsequent processing and training of the model.

[0049] Specifically, the assessment of the health status is processed based on a deep learning model. In this embodiment, the deep learning model can be a multi-layer perceptron (MLP, Multi-Layer Perceptron), and its basic structure includes an input layer, several hidden layers, and an output layer.

[0050] In one embodiment, the number of neurons in the input layer is the same as the data dimension. For example: User basic data (such as height, weight, etc.) occupies 5 neurons; Gene data (such as metabolic-related SNP locus characteristics) occupies 20 neurons; Lifestyle data (such as diet preferences, exercise intensity) occupies 10 neurons.

[0051] The design of the hidden layer is based on the model training requirements and usually includes 2 - 3 layers. The number of neurons in each layer can be 2 to 3 times the number of neurons in the input layer. The output layer of the model generates multiple health assessment results. For example: Whether the BMI index is normal; Whether there is a risk of chronic diseases; The level of the current fat metabolism ability.

[0052] As an option, the calculation formula of the BMI index is as follows: where W represents the weight (unit: kg), H represents the height (unit: m).

[0053] Formula explanation: BMI is a commonly used weight assessment indicator: BMI < 18.5: Underweight; 18.5 ≤ BMI < 24.9: Normal; 25 ≤ BMI < 29.9: Overweight; BMI ≥ 30: Obese.

[0054] In some embodiments, the model combines the user's BMI index with gene metabolism data to further predict their metabolic efficiency. Specifically, the classification prediction of metabolic efficiency can be achieved through a logistic regression model. The goal of the model is to output the category of the user's metabolic efficiency (such as "high", "medium", "low") based on gene characteristics (such as the FTO gene locus) and BMI value.

[0055] To achieve the prediction of disease risk, a genetic algorithm is introduced in this embodiment. The genetic algorithm simulates the process of natural evolution and optimizes the evaluation results of the user's health risk through operations such as selection, crossover, and mutation.

[0056] In one possible implementation, the specific implementation of the genetic algorithm is as follows: Population initialization: Randomly generate an initial solution set of health risk assessment parameters; Fitness function: Define a fitness function for health risk based on the user's genetic information and historical health data. For example, the calculation formula for hypertension risk can be: R = w 1 ×Age + w 2 ×SNP rs699 + w 3 ×Lifestyle Where, R represents the hypertension risk value, w 1 , w 2 , w 3 are the weight parameters for age, the rs699 gene locus characteristic, and lifestyle data respectively; SNP rs699 represents the user's gene characteristic value, which comes from the single nucleotide polymorphism locus rs699 in gene detection and usually takes values of 0 (unmutated), 1 (single-sided mutation), or 2 (double-sided mutation).

[0057] Lifestyle represents the user's lifestyle score, which is comprehensively scored based on habits such as diet and exercise and usually has a value range of [0, 100].

[0058] Crossover and mutation: Generate the next generation of solutions through crossover operations and introduce a small amount of random mutation to increase the diversity of the solutions.

[0059] Termination condition: When the evaluation result of the risk prediction model is stable or reaches the expected accuracy, stop the iteration.

[0060] In this embodiment, the output form of the evaluation result is a health report. This report is presented to the user in the form of charts and text, including: A quantitative score of the current health status (e.g., the metabolic score is 85 points); Possible health risks (e.g., there is a risk of mild fatty liver); Health improvement suggestions (e.g., it is recommended to increase the daily exercise time by 30 minutes).

[0061] In some embodiments, the health report will also compare the user data with other people in the same age group or within the same BMI range. Specifically, a reference interval based on population characteristics can be generated through big data analysis. For example, a user's metabolic score is 80 points, indicating that they rank in the top 20% among people of the same age group.

[0062] Through the above steps, the health assessment model can comprehensively and accurately analyze the user's health status and potential risks. These assessment results will directly serve as the input basis for generating the next personalized dietary plan, providing scientific and highly targeted health management suggestions for the user.

[0063] The generation of the personalized dietary plan is based on the health assessment results. The system generates a personalized dietary plan through machine learning algorithms: Analysis of nutritional requirements: Calculate the daily recommended energy intake and the proportion of macronutrients (e.g., 40% carbohydrates, 30% protein, 30% fat) according to the user's gender, age, activity level, and the evaluated health goals (such as weight loss, muscle enhancement).

[0064] Food recommendations: Screen ingredients that meet the user's nutritional requirements from the dynamically updated food database built into the system to generate recommended recipes for three meals a day. The food database includes more than 1,000 common ingredients, recording their nutritional components, calories, and common cooking methods.

[0065] In the foregoing steps, the system has fully grasped the user's personalized health status and potential health risks through data collection and health assessment. Based on these assessment results, the task of this step is to generate a personalized dietary plan exclusive to the user. The formulation of the dietary plan is not simply the generation of a fixed template, but is completed through dynamic calculation, data matching, and multi-layer algorithm optimization. This process needs to fully consider factors such as the user's health needs, eating habits, genetic information, etc., and combine the results of big data analysis and user preferences to ensure the scientific nature and personalization of the dietary plan.

[0066] In this embodiment, the core of meal plan generation lies in the use of dynamic adjustment algorithms and the comprehensive response to personalized needs. The system first extracts key indicators based on the results of health assessments, including the user's basal metabolic rate, daily calorie requirements, and macronutrient distribution ratios. Subsequently, through the food ingredient recommendation and recipe generation module, a meal plan suitable for the user's actual needs is developed.

[0067] Calculation of Daily Calorie Requirements Generally, the user's daily calorie requirements (TDEE, Total Daily Energy Expenditure) can be calculated using the following formula: TDEE = BMR × PAL Where, BMR is the user's basal metabolic rate, in kcal / day; PAL is the physical activity level factor, which is determined according to the user's exercise intensity. For example, for light activity, it is 1.2; for moderate activity, it is 1.55; for high-intensity activity, it is 1.9.

[0068] As an option, the system can also dynamically adjust the TDEE. For example, for the goal of fat loss, the system will recommend a daily calorie intake of 80%-90% of the TDEE to achieve a calorie deficit. At the same time, if the health assessment shows a risk of muscle loss for the user, it is recommended to have a daily calorie intake slightly higher than the TDEE to support protein synthesis.

[0069] Macronutrient Distribution Specifically, the proportions of the three major macronutrients (carbohydrates, proteins, and fats) in the meal plan are dynamically calculated based on the user's health goals. In one possible implementation: If the user's goal is fat loss, the recommended ratio is 40% carbohydrates, 35% protein, and 25% fat; If the user's goal is muscle gain, the recommended ratio is 50% carbohydrates, 30% protein, and 20% fat.

[0070] In some embodiments, the calculation formula for protein requirements is as follows: P = W × R Where, P represents the daily protein requirement (unit: grams); W represents the user's weight (unit: kilograms); R represents the protein intake coefficient, generally ranging from 1.2 to 2.0, which is adjusted according to the user's exercise intensity and health goals. For general healthy users: 1.2 - 1.5 g / kg. For users with the goal of muscle gain: 1.6 - 2.0 g / kg.

[0071] Food Ingredient Recommendation As a possible implementation, the system will match suitable ingredients in the ingredient database based on the user's calorie and nutritional requirements. The database records the nutritional components of thousands of common ingredients, including the calories, protein content, carbohydrate content, and fat content per 100 grams of ingredients. For example: 100 grams of chicken breast: calories 165 kcal, protein 31 g, carbohydrates 0 g, fat 3.6 g; 100 grams of broccoli: calories 34 kcal, protein 2.8 g, carbohydrates 6.6 g, fat 0.4 g.

[0072] Generally, the system will give priority to recommending ingredients that the user has marked as "preferred". For example, if the user prefers low-fat meats, chicken breast and fish will be recommended first. At the same time, if the user has an allergy record for certain ingredients (such as nuts), the system will automatically exclude the relevant ingredients.

[0073] Diet Plan Optimization The diet plan is not fixed, but can be dynamically optimized according to the user's real-time feedback and data. In this embodiment, the system first generates an initial diet plan through an algorithm, and then optimizes it according to the following rules: Nutritional balance check: The system will check whether the diet plan meets the target ratio of macronutrients. If a certain nutrient exceeds the standard or is insufficient, the types or portions of ingredients will be adjusted. For example, if the fat intake exceeds the target value, the system may recommend reducing nuts or oil-based ingredients.

[0074] Ingredient feasibility adjustment: Consider the user's actual eating habits and cooking conditions. For example, if the user does not have a microwave oven, the system will avoid recommending recipes that require microwave heating.

[0075] User satisfaction feedback: The system optimizes recipe recommendations based on the user's historical ratings. For example, if the user prefers Chinese cuisine, the system will give priority to recommending ingredient combinations that conform to Chinese cuisine habits.

[0076] Recipe Generation In a possible implementation, the diet plan generated by the system will be specific to the recipes for each meal, including a detailed ingredient list and cooking instructions. For example: Breakfast: 200 grams of oatmeal, 2 eggs, 1 banana; Lunch: 150 grams of chicken breast, 100 grams of broccoli, 150 grams of sweet potato; Dinner: 100 grams of salmon, 100 grams of quinoa, 100 grams of spinach.

[0077] To enhance the user experience, the system also supports adjusting the recipes according to the calorie requirements of each meal. For example, when the user needs to reduce the calorie intake at lunch, the system may reduce the portion of sweet potato and increase the proportion of broccoli.

[0078] Optimization Mechanism of Recommendation Algorithm In some embodiments, the recommendation algorithm uses collaborative filtering technology to predict users' recipe preferences. The collaborative filtering model compares a user's historical selections with those of other users to discover potential preferences. For example: If user A likes chicken breast and has selected the recommended recipes multiple times, and user B also likes chicken breast, the system may recommend high-rated recipes of user A to user B.

[0079] This collaborative filtering model can be described by the following formula: Where, S AB represents the similarity between user A and user B; R A,i and R B,i respectively represent the ratings of user A and user B for item i; i is the index of the item, representing the object recommended in collaborative filtering.

[0080] Through the above steps, the system generates a highly personalized meal plan. This plan not only meets the users' health goals but also takes into account eating habits and practical conditions, ensuring practicality and scientificity. All meal plans can be adjusted in real time to form a closed-loop optimization.

[0081] Real-time Monitoring and Dynamic Optimization Users use a smart scale to weigh ingredients and identify cooking methods, and the system uploads data to the server in real time.

[0082] The server compares the difference between the actual intake of the user and the recommended plan and adjusts the subsequent meal plan through a reinforcement learning model. For example, when it detects that the user's protein intake is lower than the recommended value, the system will preferentially increase high-protein ingredients (such as chicken breast, fish) in the next recommendation.

[0083] Optimize the recommendation algorithm by combining user feedback (such as satisfaction ratings).

[0084] In the foregoing steps, the system has generated a personalized meal plan for the user. However, this does not mean static execution of the meal plan. The actual ingredient intake, cooking methods, and related behaviors of the user in real life may deviate from the plan. Therefore, in step 4, real-time monitoring of the user's eating behavior becomes the core link. By using a smart scale and other IoT devices, the system can dynamically collect the actual food intake data of the user, and analyze and compare it with the foregoing meal plan. This link aims to establish a data closed-loop and provide direct input for subsequent optimization.

[0085] In this embodiment, the system uses a variety of devices and technical means to monitor the user's dietary behavior in real time. Among them, the smart scale is the core hardware device. This device is not only used to measure the weight of ingredients, but also can identify the cooking method and upload the data to the cloud. Through real-time monitoring, the system can more accurately understand the user's dietary behavior and ensure the implementation and adjustment of the dietary plan.

[0086] Data Acquisition and Processing of Smart Scale Generally, when the user prepares meals, the ingredients are placed on the smart scale for weighing. The smart scale collects the weight data of the ingredients through high-precision sensors and marks them in combination with the user ID. As a possible implementation, the accuracy of the weight sensor can reach 1 gram, ensuring that the measurement error of small-portion ingredients (such as seasonings) is controlled within ±0.1%.

[0087] In some embodiments, the smart scale not only collects weight data, but also identifies the types of ingredients. For example, through the integrated optical sensor, the smart scale can capture the color and texture information of the ingredients and analyze them using a pre-set image classification model. The output of the model includes the type of ingredient and the estimated calorie value. For example: If the identified ingredient is chicken breast (150 grams), the system will record the total calorie of this ingredient as 165 kcal; If the identified ingredient is broccoli (100 grams), the recorded calorie is 34 kcal.

[0088] Identification of Cooking Methods Specifically, the cooking method has a significant impact on the nutritional components of food. Therefore, identifying the user's cooking method is another important content of this embodiment. In a possible implementation, the smart scale can capture the heating time and temperature change curve of the ingredients through the built-in temperature sensor and time sensor. The system matches these data with the pre-set cooking mode database to identify the specific cooking method. For example: When the temperature rises slowly and stabilizes at 100 °C, the system determines it as steaming; When the temperature rises rapidly and fluctuates between 160 °C and 200 °C, the system determines it as frying and stir-frying.

[0089] In some embodiments, the smart scale can also work in cooperation with other smart devices (such as a smart stove) in the user's kitchen. For example, when the user is stir-frying, the gas consumption data of the stove is synchronously transmitted to the cloud and associated with the ingredient data for analysis.

[0090] Data Upload and Cloud Processing In this embodiment, all the collected data will be uploaded to the cloud in real time through the wireless communication module. As an option, the TLS protocol is used for encryption during data upload to ensure the security of transmission. The uploaded data includes the following content: The type, weight, and calories of each ingredient; The cooking method of each dish and the corresponding nutritional changes; The deviation between the user's actual intake and the dietary plan.

[0091] After receiving the data, the cloud will classify and structure it. For example, if the fat content actually ingested by the user exceeds the planned value, the system will mark this abnormal data and trigger an optimization process.

[0092] Dynamic adjustment of the dietary plan In this embodiment, the system dynamically adjusts the user's dietary plan based on the real-time collected data. For example: When the amount of protein actually ingested by the user is lower than the planned value, the system will give priority to recommending high-protein ingredients (such as fish and lean meat) in the next meal; If it is detected that the user has not achieved the planned calorie target for consecutive days, the system may reduce the difficulty of the health goal (such as reducing the calorie deficit requirement) to improve the executability of the plan.

[0093] In some embodiments, the system also supports an instant feedback function. For example, after the user ingests a certain high-calorie ingredient, the system will remind the user to reduce the calorie intake of other meals on the same day.

[0094] Integration of user feedback The user's real-time feedback is also an important basis for adjusting the dietary plan. In this embodiment, the user can evaluate the daily dietary plan and actual diet through the client. For example, the user can mark certain recipes as "difficult to operate" or "preferred". The system uses this feedback data as an important input for optimizing the dietary plan. For example: If a user marks "the breakfast recommendation is too complicated" three times in a row, the system will automatically recommend a simpler breakfast combination, such as fruit and oatmeal; If the user expresses dissatisfaction with a certain type of ingredient multiple times (such as a certain vegetable), the system will reduce the priority of this ingredient in the recommendation.

[0095] Data analysis and closed-loop optimization As a possible implementation, after analyzing the user's actual eating behavior, the system uses reinforcement learning technology to optimize the dietary recommendation algorithm. The reward function of reinforcement learning is designed as follows: Among them, R is the reward value; n is the total number of types of nutrients (unitless), such as protein, carbohydrates, fat, etc.; A i is the amount of nutrient i actually ingested by the user; T i It is the target value of the corresponding nutrient in the dietary plan.

[0096] When the actual intake value approaches the target value, the reward value approaches zero, and the algorithm regards the current recommendation as successful; otherwise, the algorithm adjusts the recommendation strategy.

[0097] Through the above implementation manner, the system can achieve real-time monitoring of the user's eating behavior and dynamic adjustment of the dietary plan. This link not only enhances the personalization of the dietary plan but also improves the execution effect of the plan through a closed-loop optimization mechanism. With the support of seamlessly connected devices and platforms, users can gradually develop healthier eating habits and obtain higher-quality nutrition management services. As part of this application, the present invention also provides a full-range personalized nutrition management system, which integrates user-side devices, smart hardware, cloud servers, and client applications to build a complete nutrition management platform from data collection to dynamic feedback optimization. The system works collaboratively through a health assessment module, a dietary plan generation module, and a data optimization module, and combines real-time user data to achieve the automation and efficiency of personalized health guidance. The system design focuses on modularity and scalability to ensure efficient and secure data processing capabilities even in large-scale deployments.

[0098] The full-range personalized nutrition management system of the present invention runs through the entire process of data collection, health assessment, dietary plan generation, and dynamic adjustment, aiming to provide an integrated nutrition management solution. The system realizes the collection, analysis, and optimization of multi-dimensional data through the collaborative work of cloud servers, user-side devices, and smart hardware, providing users with comprehensive and real-time health management services. Specifically, the system structure consists of a cloud computing center, a user-side application program, and smart hardware devices. Each module is connected through a secure communication protocol to form a closed loop of data collection, processing, and feedback.

[0099] In this embodiment, the system consists of three major modules, namely a cloud server, a user-side device, and smart hardware. Each module is connected and works collaboratively through wireless communication technology.

[0100] Cloud server As the core computing and storage center of the system, the cloud server is mainly responsible for health assessment, dietary plan generation, and data optimization processing. Generally, the cloud server consists of multiple functional modules, including a health assessment module, a dietary plan generation module, a data storage module, and a data optimization module.

[0101] In a possible implementation, the health assessment module runs on a high-performance computing node, which processes the data uploaded by the user based on a pre-trained deep learning model to generate a health status assessment result. For example, the deep learning model can receive the user's basic health data, genetic feature data, and lifestyle data as inputs through a multi-layer perceptron architecture, and output the BMI status, fat metabolism score, and chronic disease risk assessment result.

[0102] The dietary plan generation module works in cooperation with an optimization algorithm through a logistic regression model. Specifically, the dietary recommendation is completed by the following steps: Through the logistic regression model, according to the user's health goals (such as fat loss, muscle gain), determine the target proportions of the three major macronutrients; Combined with the user's dietary preferences, screen suitable ingredients from the ingredient database; Generate a daily dietary plan that meets the user's calorie and nutritional requirements through an integer linear programming algorithm.

[0103] As an option, the cloud server adopts a distributed storage architecture, and shards the user's health data and dietary plans and stores them on different nodes to ensure data security and efficiency. The data will be encrypted before storage, using the AES-256 encryption algorithm to ensure that the user's privacy is not leaked.

[0104] User terminal device The user terminal device mainly exists in the form of a mobile application, and its functions include user data input, display of health assessment reports, viewing of dietary plans, and collection of user feedback. In this embodiment, the user terminal device provides an instant health management service for the user through real-time interaction with the cloud server.

[0105] Specifically, the user enters personal basic health data (such as height, weight, age) in the client application, uploads a gene detection file (such as in CSV or TXT format), and completes a lifestyle questionnaire. As a possible implementation, the user terminal device will perform local verification on the entered data, such as checking the range of the user's weight value to avoid abnormal data affecting subsequent evaluations.

[0106] The user terminal device also undertakes the function of displaying the health assessment result. For example, when the user completes data upload, the client will present the assessment result in a graphical interface, including the user's current health status score, chronic disease risk level, and recommended health goals.

[0107] In some embodiments, the client device integrates a virtual assistant module. Specifically, based on natural language processing technology and sentiment analysis algorithms, this module can answer users' health questions. For example, when a user asks "What is the healthiest thing to have for dinner today?", the virtual assistant will recommend a suitable recipe according to the daily meal plan and the user's preferences.

[0108] Smart hardware Smart hardware is a key part of data collection in the system. In this embodiment, smart hardware mainly refers to a smart scale, which is used to collect data on users' food intake and cooking methods.

[0109] Generally, a smart scale consists of a high-precision weight sensor, an optical recognition module, a data transmission module, and an embedded processor. The weight sensor can measure the weight of food ingredients in the range from 0.1 grams to 10 kilograms, with an error of no more than ±0.1 grams. The optical recognition module captures images of food ingredients through a camera and uses an embedded convolutional neural network (CNN) model to classify the types of food ingredients. For example, when a user places a raw chicken breast on the scale, the smart scale will identify the food ingredient as a chicken breast, with a weight of 150 grams, and upload its corresponding calorie value (165 kcal).

[0110] In a possible implementation, the smart scale can also detect the heating situation of food ingredients through a built-in temperature sensor to determine the cooking method. For example, when the temperature rises above 200°C, the system can determine that the food ingredients are being fried.

[0111] The smart scale communicates with the cloud server via WiFi or Bluetooth. Before data transmission, the data will be encrypted through the TLS protocol to ensure that users' privacy is not stolen.

[0112] Communication and security mechanisms In this embodiment, the various modules of the system achieve data interaction through wireless communication technologies, and the communication protocols support WiFi, Bluetooth, and cellular networks. In a possible implementation, all data is encrypted using the TLS1.3 protocol during transmission to avoid leakage of sensitive information.

[0113] In addition, the cloud server adopts a hierarchical permission management mechanism for user data. Specifically, users can only access and modify their own data, and administrators can only perform data maintenance operations when authorized. The authentication mechanism supports two-factor authentication (such as password + SMS verification code) to further enhance the security of the system.

[0114] Data optimization and closed-loop management In this embodiment, the data optimization module continuously adjusts the recommendation algorithm based on the user's actual dietary behavior data and dietary feedback data. For example, when the protein intake in the user's actual diet is lower than the planned target, the system will automatically adjust the dietary plan for the next meal and increase the recommendation weight of high-protein ingredients.

[0115] In some embodiments, the data optimization module operates based on a reinforcement learning algorithm. Its reward function is designed as follows: Wherein, R represents the reward value; n is the total number of nutrient types (unitless), such as protein, carbohydrates, fat, etc.; A i represents the quantity of nutrient i actually ingested by the user; T i represents the target value of the corresponding nutrient in the dietary plan.

[0116] Through real-time adjustment, the system can gradually optimize the personalization level of the dietary plan and provide users with higher-precision nutrition management services.

[0117] Through the above implementation manners, the all-round personalized nutrition management system of the present invention realizes the closed-loop operation of data collection, processing, and feedback. The cloud server, user terminal device, and intelligent hardware in the system each perform their own functions and complement each other, and can comprehensively meet the health management needs of users, providing users with scientific, real-time, and dynamic personalized dietary recommendations and health management services. As part of this application, the present invention also provides a smart scale. As the core hardware device for user data collection, the smart scale has functions of high-precision weight measurement, cooking mode recognition, and data encrypted transmission. The smart scale accurately collects the user's food intake data through the built-in sensors and data processing module and uploads it in real time, providing key support for the generation of personalized dietary plans. Its design focuses on the combination of hardware and algorithms, not only improving the efficiency of data collection but also greatly reducing human errors.

[0118] In this embodiment, the smart scale mainly includes the following hardware components: Weight sensor: used to measure the weight of food ingredients; Optical recognition module: used to identify the types of food ingredients; Temperature sensor: used to detect the temperature change during the cooking process; Data transmission module: realizes data uploading and communication; Embedded processor: used to run food ingredient recognition and data processing algorithms.

[0119] Generally, the weight sensor of the smart scale can achieve an accuracy of ±0.1 grams, covering a weight measurement range from 0.1 grams to 10 kilograms, meeting the measurement requirements of various food ingredients in daily home cooking. For example, when a user places a raw chicken breast (weighing about 150 grams) on the scale, the sensor will immediately measure the weight and transmit the data to the embedded processor.

[0120] As a possible implementation, the optical recognition module consists of a high-resolution camera and an embedded image processing chip. After the user places the food ingredient on the smart scale, the camera captures the image of the food ingredient, and analyzes the image through a Convolutional Neural Network (CNN) model, outputting the category of the food ingredient and a preliminary estimated calorie value. For example: If the food ingredient placed is an apple, the system will identify it as an apple and estimate the calorie to be 52 kcal / 100 grams; If the food ingredient placed is a raw steak, it will be recorded as "steak", and the user will be prompted to select a specific cooking method (such as frying, boiling).

[0121] Specifically for cooking method recognition, the cooking method will significantly affect the nutritional composition change of the food ingredient. Therefore, the smart scale is designed with a temperature sensor for detecting the cooking process and a time recording module. In a possible implementation, the temperature sensor monitors the temperature change curve on the surface of the food ingredient in real time and matches it with the built-in cooking mode database in the system. The common cooking methods in the database include steaming, boiling, frying, baking, etc.

[0122] In some embodiments, the system determines the cooking method according to the following rules: If the temperature rises slowly and stabilizes at about 100°C, it is determined to be steaming or boiling; If the temperature rises rapidly and fluctuates between 180°C and 200°C, it is determined to be stir-frying; If a temperature as high as 250°C is detected, the system may determine it to be baking or frying.

[0123] To further improve the recognition accuracy, the smart scale can be linked with a smart stove. For example, if the stove detects a large amount of oil and a long duration of high temperature, the system will mark this cooking method as "frying".

[0124] Data Processing and Storage In this embodiment, the embedded processor runs a multi-threaded data processing algorithm. Generally, the data collected by the sensor will first undergo formatting processing by the processor. For example, the weight, type, and cooking method of the food ingredient are stored as structured data respectively. After processing, the data will be uploaded to the cloud server through the data transmission module.

[0125] As a possible implementation, the smart scale communicates with the cloud via WiFi or Bluetooth. When transmitting data, the TLS protocol is used for encryption to ensure the security of users' privacy information during the transmission process.

[0126] In some embodiments, the smart scale also supports the local storage function. When the network connection is unavailable, the data will be temporarily stored in the built-in non-volatile memory and automatically uploaded after the network is restored. The capacity of the memory is generally designed to be 128MB, which can store the user's diet data for 1-2 weeks.

[0127] Multi-user management and marking function In a possible implementation, the smart scale supports multi-user management. Specifically, before using the smart scale, the user needs to scan the QR code or enter the personal ID through the mobile application for binding. The system will assign a unique identifier to each data record according to the bound user ID to ensure that multi-user data is not confused.

[0128] In addition, the smart scale provides a marking function. The user can manually input or select the cooking method to correct the results recognized by the system. For example, when the user selects a specific complex cooking method (such as braised), the system will synchronize the marking information to the cloud.

[0129] Calculation of nutritional components of ingredients In this embodiment, the database built into the smart scale records the nutritional components of common ingredients (such as calories, protein, carbohydrates, and fat content per 100 grams). When the user places the ingredients, the system will calculate the ingested nutritional components in real time according to the weight and type of ingredients. For example: When the user places 100 grams of chicken breast, the system will record 165 kcal of calories, 31 g of protein, 3.6 g of fat, and 0 g of carbohydrates; if the user places 150 grams of cooked rice, the system will record 207 kcal of calories, 4.5 g of protein, 45 g of carbohydrates, and 0.3 g of fat.

[0130] As an option, the system can also calculate the total calories of the whole meal in real time. For example, if the user weighs chicken breast (150 grams) and broccoli (100 grams) on the scale in sequence, the system will display the total calories as 199 kcal.

[0131] System linkage and expansion function In some embodiments, the smart scale can not only run independently but also work in coordination with other smart devices. For example, the smart scale can be linked with the mobile client to synchronize the data to the dietary record module of the client in real time. The user can view the daily ingredient intake situation in the client and understand the distribution of the three major macronutrients through the visual chart.

[0132] As a possible extended function, the smart scale also supports a voice interaction module. For example, when the user uses the smart scale to weigh, the voice assistant will real-time announce the calorie value of the food ingredients and the current total calories, helping the user to instantly grasp the diet situation.

[0133] Ease of use In this embodiment, the design of the smart scale fully considers the ease of use of the user. For example: The tabletop of the smart scale is made of waterproof material, which is convenient for the user to clean during the cooking process; The display screen uses a large-size LCD screen, which can clearly display the food ingredient information and the data upload status; Anti-slip pads are installed at the bottom of the scale body to ensure the stability of the scale during use.

[0134] Through the above implementation methods, the smart scale realizes the full-process automatic operation from data collection to cloud upload. Its high-precision data collection, versatility and good user experience provide important technical support for the all-round personalized nutrition management system of the present invention.

[0135] As part of this application, the present invention also provides a cloud server for all-round personalized nutrition management. As the data processing center of the system, the server is responsible for running the health assessment model, the dietary plan generation algorithm and the data optimization module, and ensuring the security of data storage and transmission. Through big data analysis and AI algorithms, the server comprehensively processes the user's health data and generates personalized suggestions, and continuously optimizes the health management ability of the system based on reinforcement learning technology. The server design adopts a distributed storage architecture and multi-level permission control, which not only ensures the efficient operation of the system, but also strictly protects the user's privacy.

[0136] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A comprehensive personalized nutrition management method, characterized by: The following steps are involved: S1. Obtain the user's multi-dimensional health data through the user-end device, including genetic information, living habits, exercise frequency and food intake information; S2. Use AI-driven health assessment models to analyze the above multi-dimensional health data and generate assessment results of the user's health status and potential risk factors; S3. generating a personalized meal plan using a machine learning algorithm based on the health assessment results; S4. Collect the user's actual food intake data and cooking method through the smart scale and upload it to the server in real time; S5. Dynamically optimize personalized meal plans based on user feedback information and real-time collected data.

2. The all-round personalized nutrition management method according to claim 1, characterized in that: The health assessment model in step S2 adopts a combination of genetic algorithms and deep learning algorithms to obtain personalized health risk analysis results based on big data training.

3. The all-round personalized nutrition management method according to claim 1, characterized in that: In step S5, the dietary recommendation strategy is dynamically adjusted by regularly collecting user usage feedback and health indicator changes using a reinforcement learning mechanism.

4. Comprehensive personalized nutrition management system, characterized by: The all-round personalized nutrition management method applied to any one of claims 1 to 3 comprises: User-end devices are used to collect multi-dimensional health data of users; Smart scales to collect data on users’ food intake and cooking methods; The cloud server includes: a health assessment module for generating health status assessment results based on user health data; a meal plan generation module for generating personalized meal plans based on health assessment results; and a data optimization module for optimizing meal plans based on user feedback and real-time collected data; Client application for interacting with users and displaying meal plans and health assessment results.

5. The all-round personalized nutrition management system according to claim 1, characterized in that: The cloud server also includes a virtual assistant module, which combines natural language processing technology and a sentiment analysis engine to push personalized health advice and recipe recommendations based on user preferences and emotional states.

6. The all-round personalized nutrition management system according to claim 1, characterized in that: The user terminal device includes a mobile application for collecting user health data and displaying personalized meal plans and health status assessment results.

7. A smart scale, characterized in that: The all-round personalized nutrition management method applied to any one of claims 1 to 3 comprises: Food weight sensor, used to measure the weight of the user's ingredients; A cooking mode recognition module, used to recognize the user's cooking method; A data transmission module for uploading food weight and cooking mode information to a cloud server in real time; Built-in processing module to format the collected data and bind it to the user ID.

8. The smart scale according to claim 1, characterized in that: The data transmission module uses an encryption protocol to encrypt the data to ensure the security of the food intake data during transmission.

9. A cloud server for all-round personalized nutrition management, characterized in that: The all-round personalized nutrition management method applied to any one of claims 1 to 3 comprises: A health assessment module is used to generate health status assessment results based on the user's multi-dimensional health data; A meal plan generation module is used to generate a personalized meal plan based on the health assessment results; Data optimization module, used to optimize meal plans based on user feedback and real-time data collection; The data storage module is used to store user health data and meal plans, and set multi-level permission control.

10. A cloud server for all-round personalized nutrition management according to claim 9, characterized in that: A distributed storage architecture is used to encrypt and store user health data, and access rights are controlled through a multi-layer authentication mechanism.

Images