Personalized nutrition bag planning method and device based on child health management, equipment and medium

Abstract

The present application relates to a kind of personalized nutrition package planning method, device, equipment and medium based on child health management, method includes: integrating child identity, growth and biochemical data constructs individualized nutrition file and global benchmark knowledge base;Based on time series growth data and daily activity data, individualized daily nutrition amount vector is calculated and generated;Real-time dietary intake vector is generated by recognizing meal image and matching knowledge base data;The above two vectors are fused, and the interaction and safety constraints of nutrients are considered, and the theoretical optimal nutrition supplement vector is calculated by optimization model;Finally, the vector is matched with the nutrition component library, and the nutrition package formula that can be directly executed is generated.The method solves the problems of static lag, data fragmentation, neglect of nutrient synergy / antagonism and theoretical scheme difficult to implement in the prior art, realizes personalized nutrition supplement from the real-time state of children, precise and safe and operable.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of data processing and health management, specifically to a method, apparatus, equipment, and medium for planning personalized nutrition packages based on children's health management. Background Technology

[0002] In the field of child health management, personalized nutritional intervention is a key means to promote normal growth and development in children and prevent nutritional deficiencies or excesses. Nutritional supplement packages, as a targeted nutritional supplement carrier, offer the core value of providing precise nutrient ratios based on a child's individual real-time health status and dietary gaps. Therefore, the accompanying personalized nutritional supplement planning method aims to generate a "tailor-made" supplementation plan for each child through systematic assessment and calculation, thereby transforming scientific nutritional recommendations into concrete and actionable daily intervention measures.

[0003] However, existing nutritional supplement planning methods still have significant limitations in practical applications. First, most methods rely on periodic health check data and static nutrient recommendations, failing to integrate dynamic information about children on a given day (such as actual activity levels and immediate dietary intake). This results in a disconnect between the recommended nutritional supplement components and children's real-time, actual needs, making it difficult to achieve a precise match of "supplementation immediately after meals." Second, the planning process often separates health assessment data from daily dietary records, lacking an automated data bridge. This makes nutrient gap calculations reliant on manual estimation, leading to insufficient accuracy. More importantly, existing methods typically consider each nutrient in isolation when calculating supplementation amounts, ignoring the complex synergistic or antagonistic effects between nutrients, potentially affecting the overall effectiveness and safety of supplementation. Finally, the generated plans often remain at the theoretical recommendation level, failing to automatically link and convert them to a specific, readily available library of basic nutritional ingredients. This results in a missing link between the "theoretical plan" and the "executable formula," reducing the plan's practicality and user compliance. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide a method, device, equipment and medium for planning personalized nutrition packages based on children's health management that can achieve real-time dynamic assessment, precise nutritional complementarity, safety, efficiency and execution.

[0005] The objective of this invention is achieved through the following solution:

[0006] In a first aspect, the present invention provides a personalized nutrition package planning method based on children's health management, comprising the following steps:

[0007] S1: Integrate the identity information, continuously recorded anthropometric data and biochemical index data from the collected raw data of child users into individualized nutrition profiles, and build a global benchmark knowledge base containing standard food nutrient component vectors and nutrient attribute rules based on a pre-set nutrition standard database.

[0008] S2: Based on the time-series growth data in the individualized nutrition profile, fit the personalized growth rate of the child user, and combine the obtained daily activity level data of the child user with the standard reference value to calculate and generate an individualized daily nutrient requirement vector covering multiple nutrients;

[0009] S3: Recognize and process food images uploaded by users through the terminal, identify food types and estimate the quality of various identified foods, query the global benchmark knowledge base to match nutritional component data, accumulate and calculate to generate a real-time dietary intake vector.

[0010] S4: Calculate the original nutritional gap based on the individualized daily nutrient requirement vector and the real-time dietary intake vector, and construct a nutritional complementarity optimization model based on the nutrient interaction rules and safety constraints in the global benchmark knowledge base to calculate the optimized nutritional supplementation plan and output the theoretically optimal nutritional supplementation vector.

[0011] S5: Match and fit the theoretically optimal nutritional supplementation vector with the pre-stored nutritional component library to solve for the specific raw material weight ratio that satisfies the nutritional target defined by the theoretically optimal nutritional supplementation vector, and generate a nutritional pack formula. The nutritional pack formula is used to indicate the ratio of nutrients that supplement the nutritional gap required by the child user on that day.

[0012] In one embodiment, S2 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0013] S21: Perform trend analysis on continuously recorded anthropometric data in individualized nutrition records, apply time series models to fit growth curves and calculate rate changes to generate personalized growth rates.

[0014] S22: Quantitatively evaluate and process the acquired activity data of child users on the same day, converting the activity duration, type and intensity into standard metabolic equivalent values ​​to generate activity level data;

[0015] S23: Calculate the deviation of the personalized growth rate from the standard growth rate of the global benchmark knowledge base and the ratio of the activity level data to the benchmark activity level of the global benchmark knowledge base as dynamic adjustment factors. Input the dynamic adjustment factors into the preset personalized nutrition requirement prediction model, and perform weighted calculation of the standard nutrient reference intake and the dynamic adjustment factors to generate an individualized daily nutrient requirement vector.

[0016] In one embodiment, S3 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0017] S31: Perform feature extraction and classification processing on the food images uploaded by users, identify the food categories in the images through convolutional neural networks, estimate the physical quality of each type of food based on image features, and generate a set of recognition results containing food labels and quality pairs;

[0018] S32: Perform component mapping processing on each food identifier in the recognition result set, query the global benchmark knowledge base to match its standard nutrient content data, and generate a nutrient vector corresponding to the identified food.

[0019] S33: Based on the nutrient component vector and the corresponding food quality, a weighted sum is performed, and the contribution value of all meals is accumulated within the daily time window to generate a real-time dietary intake vector.

[0020] In one embodiment, S4 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0021] S41: Perform gap analysis on the individualized daily nutrient requirement vector and the real-time dietary intake vector, compare each element to determine the difference between the immediate requirement and the intake of each nutrient, and generate the original nutrient gap.

[0022] S42: Extract and parameterize the nutrient attribute rules in the global benchmark knowledge base, read the synergistic or antagonistic relationships between nutrients, the tolerable upper intake threshold, and the unit nutrient cost coefficient, and generate a set of rule parameters to describe nutrient interactions, safety limits, and economic constraints.

[0023] S43: Multi-objective fusion modeling is performed based on the original nutritional gap and the set of rules and parameters. The original nutritional gap is used as the baseline objective. At the same time, the synergistic and antagonistic relationship data in the set of rules and parameters are introduced to adjust the supplementation ratio of various nutrients. The safety and cost constraints are superimposed by the tolerable maximum intake threshold and the unit nutrient cost coefficient extracted from the set of rules and parameters, and a nutritional complementarity optimization model under multiple constraints is constructed.

[0024] S44: Iteratively solve the nutritional complementarity optimization model. Combine the preset sequential quadratic programming algorithm to iteratively adjust and evaluate the candidate nutritional supplementation schemes under the dual constraints of safety and cost. When the rate of change of the overall optimization objective function value corresponding to the candidate nutritional supplementation scheme is less than the preset convergence threshold during the iteration process, it is determined that it has converged to the optimal interval, and the current candidate scheme is output as the theoretically optimal nutritional supplementation vector.

[0025] In one embodiment, the expression of the sequential quadratic programming algorithm for a personalized nutrition package planning method based on children's health management provided by the present invention is as follows:

[0026]

[0027]

[0028]

[0029] Where x represents the planned supplementation amount of all nutrients to be optimized. The vector formed, i.e. ,in This represents the amount of the i-th nutrient to be supplemented. Let be the nutrient replenishment vector at the k-th iteration. At the current iteration point The search direction vector is to be found at the location. To describe the local curvature of the objective function at the k-th iteration, The objective function f(x) of the original problem is located at point... The gradient vector at that point, Let i be the i-th equality constraint function, and let its set index be . , Let I be the j-th inequality constraint function, and let I be its set index.

[0030] In one embodiment, S5 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0031] S51: Initialize and characterize the pre-stored nutrient component library to ensure that each basic nutrient raw material is associated with a component feature vector defined based on the content of each nutrient in the basic nutrient raw material, and generate a standardized component feature matrix.

[0032] S52: Match and fit the theoretical optimal nutrient supplementation vector with the component feature matrix, and transform it into a linear programming problem with the goal of minimizing the ratio error and the constraint of non-negativity of raw materials. This will generate a raw material weight ratio scheme that meets the requirements of the theoretical optimal nutrient supplementation vector and has the best cost.

[0033] S53: Format and generate instructions for the raw material weight ratio scheme, converting it into standardized text and lists containing raw material names, precise mass, mixing order and preparation instructions, and generating an executable nutrition pack formula.

[0034] Secondly, the present invention provides a personalized nutrition package planning device based on children's health management, which is equipped with the following modules:

[0035] The nutrition profile building module is used to integrate the identity information, continuously recorded anthropometric data and biochemical index data in the original data of collected children users into an individualized nutrition profile and build a global benchmark knowledge base containing standard food nutrient component vectors and nutrient attribute rules based on a pre-set nutrition standard database.

[0036] The daily nutrient requirement calculation module is used to fit the personalized growth rate of children based on the time-series growth data in the individualized nutrition profile, and to calculate and generate an individualized daily nutrient requirement vector covering multiple nutrients by combining the obtained activity level data of the children on that day with the standard reference value.

[0037] The dietary intake calculation module is used to identify and process food images uploaded by users through the terminal, identify food types and estimate the quality of each identified food, query the global benchmark knowledge base to match nutritional component data and accumulate calculations to generate a real-time dietary intake vector.

[0038] The nutrition supplementation optimization module is used to calculate the original nutritional gap based on the individualized daily nutrient requirement vector and the real-time dietary intake vector, and to construct a nutritional complementarity optimization model based on the nutrient interaction rules and safety constraints in the global benchmark knowledge base, and calculate the optimized nutritional supplementation plan, outputting the theoretically optimal nutritional supplementation vector.

[0039] The nutrition pack formula generation module is used to match and fit the theoretically optimal nutrition supplementation vector with a pre-stored nutrition component library, solve for the weight ratio of specific raw materials that meet the nutritional goals defined by the theoretically optimal nutrition supplementation vector, and generate a nutrition pack formula. The nutrition pack formula is used to indicate the ratio of nutrients to supplement the nutritional gaps required by child users on that day.

[0040] Thirdly, this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the above-mentioned personalized nutrition package planning methods based on children's health management.

[0041] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned personalized nutrition package planning methods based on children's health management.

[0042] In summary, the personalized nutrition package planning method based on children's health management provided in this application systematically integrates children's static records and dynamic data, and constructs a standard knowledge base, thereby laying a reliable data foundation for accurate nutrition calculation. Based on time-series growth data and personalized demand prediction of real-time activity levels, it achieves a leap from static general recommendations to dynamic individualized calculations, ensuring a close alignment between nutritional goals and children's actual physiological states. Utilizing image recognition technology to automatically analyze dietary intake effectively solves the problem of the disconnect between health data and dietary records, providing accurate input for real-time gap calculation. By introducing nutrient interaction rules and multiple safety constraints to construct an optimization model, it ensures that the generated supplementation plan not only fills quantitative gaps but also conforms to the inherent laws of nutrition and safety boundaries, thus avoiding the low absorption efficiency or potential risks that may arise from the simple superposition of nutrients in traditional methods. Automatic matching and fitting of theoretical optimization results with a physical raw material library directly transforms digital solutions into executable formulas containing specific raw material ratios and operational instructions, bridging the "last mile" from scientific calculation to actual execution, significantly improving the practicality, safety, and user compliance of personalized nutrition interventions.

[0043] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0044] Figure 1 A flowchart illustrating a personalized nutrition package planning method based on children's health management, provided for an embodiment of this application;

[0045] Figure 2 This is a schematic diagram of a personalized nutrition package planning device based on children's health management, provided as another embodiment of this application. Detailed Implementation

[0046] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0048] In one embodiment, such as Figure 1As shown, a personalized nutrition package planning method based on children's health management is provided. This embodiment illustrates the method applied to a terminal, but it is understood that the method can also be applied to a server, or to a device including both a terminal and a server, and implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0049] S1: Integrate the identity information, continuously recorded anthropometric data and biochemical index data from the collected raw data of child users into individualized nutrition profiles, and build a global benchmark knowledge base containing standard food nutrient component vectors and nutrient attribute rules based on a pre-set nutrition standard database.

[0050] Specifically, the system receives raw data collected from child users. This raw data includes identity information, continuously recorded anthropometric data, and biochemical indicator data. The system classifies and associates these three types of data, using identity information as the core index to distinguish the individual characteristics of different child users. The system stores the continuously recorded anthropometric data in chronological order according to the collection time, ensuring that the data can fully reflect the dynamic changes in the child's growth and development. The system binds the biochemical indicator data with the anthropometric data at the corresponding collection time point, forming a structured data set containing identity information, time information, vital sign information, and biochemical information. This structured data set is defined as an individualized nutrition profile.

[0051] During the construction of individualized nutrition profiles, the system simultaneously builds a global benchmark knowledge base. The system accesses a pre-set nutrition standard database, which contains relevant content on children's dietary nutrient references published by authoritative institutions and relevant content on standard food nutrient components. The system converts the relevant content on standard food nutrient components into vector form, with each food corresponding to a nutrient component vector. Each component of the vector corresponds to the composition of various macronutrients and micronutrients in the food. Based on relevant nutritional research conclusions, the system extracts nutrient attribute rules, which cover the synergistic and antagonistic effects between nutrients. The system integrates the standard food nutrient component vectors and nutrient attribute rules to form a global benchmark knowledge base.

[0052] S2: Based on the time-series growth data in the individualized nutrition profile, fit the personalized growth rate of the child user, and combine the obtained daily activity level data of the child user with the standard reference value to calculate and generate an individualized daily nutrient requirement vector covering multiple nutrients.

[0053] Specifically, the system retrieves time-series growth data stored in the individualized nutrition record. This data originates from continuously recorded anthropometric data. The system extracts all time-series growth data within a preset time range for processing. A fitting algorithm is used to analyze the extracted data, and a curve reflecting the child's growth trend is obtained through algorithmic calculation. Based on this curve, the system derives the child's personalized growth rate, which reflects the individual child's growth pattern and developmental characteristics at the current stage. The system also acquires the child's daily activity level data, which includes various information related to the child's daily activities. This data is compared with standard reference values ​​stored in a global baseline knowledge base, and an activity level correction coefficient is calculated based on this comparison.

[0054] Furthermore, the system combines personalized growth rate with activity level correction coefficients to adjust the baseline nutritional requirements stored in the global baseline knowledge base. The adjustment process strictly follows the relevant rules in the knowledge base to ensure that the adjustment results meet the individual growth and development needs of children. Based on the adjusted nutritional requirements, the system generates an individualized daily nutritional requirement vector. This vector contains the requirements for multiple nutrients, and the dimensions of the vector are consistent with the types of nutrients defined in the global baseline knowledge base. Each component of the vector corresponds to the daily requirement of one nutrient.

[0055] S3: Recognize and process food images uploaded by users through the terminal, identify food types and estimate the quality of each identified food, query the global benchmark knowledge base to match nutritional component data, accumulate and calculate to generate a real-time dietary intake vector.

[0056] Specifically, the system receives food images uploaded by users via their terminals. These images are taken and uploaded by users after children have eaten. The system preprocesses the received food images to improve their clarity and recognizability, ensuring accuracy for subsequent identification. Preferably, the system uses a deep learning object detection algorithm to analyze the preprocessed food images. The algorithm extracts the feature information of the food in the image and matches it with a food feature database stored in the system to identify the various food types in the image. The system then estimates the quality of each identified food item. This estimation is based on the food's morphological features and size information in the image, combined with relevant information from reference objects stored in the system, to ensure that the estimation results reflect the actual amount of food consumed. The system uses the identified food types and estimated food quality as query conditions, searches the global benchmark knowledge base, and matches the corresponding nutrient vector for each food. The system accumulates the nutrient vectors of all foods, and the calculation process is carried out according to the correlation between the nutrient components and corresponding quality of each food. The accumulated result forms a real-time dietary intake vector. The dimension of this vector completely corresponds to the dimension of the individualized daily nutrient requirement vector. Each component represents the actual intake of this type of nutrient by the child user on that day.

[0057] S4: Calculate the original nutritional gap based on the individualized daily nutrient requirement vector and the real-time dietary intake vector, and construct a nutritional complementarity optimization model based on the nutrient interaction rules and safety constraints in the global benchmark knowledge base to calculate the optimized nutritional supplementation plan and output the theoretically optimal nutritional supplementation vector.

[0058] Specifically, the system retrieves the individualized daily nutrient requirement vector and the real-time dietary intake vector. The system calculates the difference between the two vectors, performing the calculation one component at a time. The resulting difference forms the initial nutrient gap vector, where each component corresponds to a preliminary gap value for a specific nutrient. This preliminary gap value represents the difference between the child user's daily requirement and intake of that nutrient. The system then retrieves nutrient interaction rules and safety constraints from the global baseline knowledge base. These rules cover synergistic and antagonistic effects between nutrients, while the safety constraints are set based on relevant nutrient intake standards. Based on these rules and conditions, the system constructs a nutrient complementarity optimization model. The model uses the initial nutrient gap vector as its base data and aims to achieve a nutrient gap filling rate that meets preset requirements and ensure that the synergistic effect of nutrients conforms to the rules. The model's constraints include that the amount of each nutrient supplemented must meet safety constraints and that the ratio between nutrients must follow the interaction rules. The system employs a corresponding algorithm to solve the nutritional complementarity optimization model. The solution process involves calculating and adjusting the variables in the model to ensure that the solution results meet the constraints set by the model. The results obtained from solving the model form a theoretically optimal nutritional supplementation vector. This vector clarifies the optimal supplementation amount for each nutrient while avoiding antagonistic effects between nutrients and strengthening synergistic effects.

[0059] S5: Match and fit the theoretically optimal nutritional supplementation vector with the pre-stored nutritional component library to solve for the specific raw material weight ratio that satisfies the nutritional target defined by the theoretically optimal nutritional supplementation vector, and generate a nutritional pack formula. The nutritional pack formula is used to indicate the ratio of nutrients that supplement the nutritional gap required by the child user on that day.

[0060] Specifically, the system retrieves a pre-stored nutrient component library, which contains information on various basic nutritional ingredients, including the nutrient composition and purity parameters of each ingredient. The system preprocesses the ingredient information in the library to ensure its accuracy and completeness. The system then matches the theoretically optimal nutrient supplementation vector with the ingredient information in the library. This matching process correlates the supplementation amount of each nutrient in the vector with the content of each nutrient in the ingredient, selecting basic nutritional ingredients that cover all nutrient types in the theoretically optimal nutrient supplementation vector. The system constructs an ingredient ratio solution model. The model uses the supplementation amount of each nutrient in the theoretically optimal nutrient supplementation vector as the target value and the weight of each basic nutritional ingredient as the variable. The model's constraints include that the ingredient combination must cover all nutrient types in the vector, the deviation of the total amount of each nutrient from the theoretical value must meet preset requirements, and the use of ingredients must comply with relevant standards.

[0061] Preferably, the system can use a linear programming algorithm to solve the raw material ratio calculation model. During the solution process, the values ​​of each variable are calculated and adjusted to ensure that the final raw material weight ratio meets the nutritional target set by the theoretically optimal nutritional supplementation vector, while taking into account the availability of raw materials and the feasibility of the ratio. Based on the solved raw material weight ratio, the system generates a nutritional pack formula. This formula clearly defines the proportions of each basic nutritional ingredient and can be directly used to guide the production and preparation of nutritional packs, achieving precise supplementation of the nutritional gaps required by children on a given day, and ensuring the nutritional needs of children during their growth and development.

[0062] In summary, the personalized nutrition package planning method based on children's health management provided in this application systematically integrates children's static records and dynamic data, and constructs a standard knowledge base, thereby laying a reliable data foundation for accurate nutrition calculation. Based on time-series growth data and personalized demand prediction of real-time activity levels, it achieves a leap from static general recommendations to dynamic individualized calculations, ensuring a close alignment between nutritional goals and children's actual physiological states. Utilizing image recognition technology to automatically analyze dietary intake effectively solves the problem of the disconnect between health data and dietary records, providing accurate input for real-time gap calculation. By introducing nutrient interaction rules and multiple safety constraints to construct an optimization model, it ensures that the generated supplementation plan not only fills quantitative gaps but also conforms to the inherent laws of nutrition and safety boundaries, thus avoiding the low absorption efficiency or potential risks that may arise from the simple superposition of nutrients in traditional methods. Automatic matching and fitting of theoretical optimization results with a physical raw material library directly transforms digital solutions into executable formulas containing specific raw material ratios and operational instructions, bridging the "last mile" from scientific calculation to actual execution, significantly improving the practicality, safety, and user compliance of personalized nutrition interventions.

[0063] In one embodiment, S2 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0064] S21: Perform trend analysis on continuously recorded anthropometric data in individualized nutrition records, apply time series models to fit growth curves and calculate rate changes to generate personalized growth rates.

[0065] Specifically, the system retrieves pre-built individualized nutritional profiles and extracts continuously recorded anthropometric data. This extracted data includes core growth indicators such as height, weight, head circumference, and chest circumference of children at different collection time points, along with the corresponding collection timestamps, collection scenario information, and operational procedures of the personnel collecting the data. The system first preprocesses the extracted anthropometric data, specifically performing timestamp continuity checks to identify any abnormal collection time intervals or time gaps. It also performs data logic consistency checks to verify the reasonable matching between different growth indicators at the same time point, removing invalid data entries with discontinuous timestamps, values ​​exceeding normal physiological ranges, or duplicate records. For missing data, the system uses linear interpolation based on the temporal trends of growth data for children of the same age group to complete the missing data, ensuring data continuity and completeness over time.

[0066] After preprocessing, the system performs trend analysis on the batch of data. A sliding window smoothing process eliminates the interference of short-term random fluctuations on the data trend. Then, time-series correlation analysis is used to extract the core trend features of the data over time, clarifying the long-term variation patterns of anthropometric data. The system applies a time-series model to fit the preprocessed anthropometric data. This model, based on the temporal characteristics of children's growth data, constructs a mapping relationship between time-dimensional variables and anthropometric index variables. By iteratively optimizing the model's core parameters, the deviation between the model output value and the actual measurement value is continuously reduced until the deviation stabilizes within a preset range, ensuring that the fitted growth curve closely matches the actual growth trajectory of children. After the growth curve is fitted, the system calculates the rate of change based on the curve. By solving the derivative of the curve at each key time node, the instantaneous growth rate of the corresponding stage is obtained, thus revealing the rate of change at different growth stages. The system integrates the instantaneous growth rate data of each stage with the long-term variation patterns obtained from trend analysis through time-series correlation, eliminating abnormally fluctuating rate data, and generating personalized growth rates that reflect the individual growth patterns of children.

[0067] S22: Quantitatively evaluate and process the acquired activity data of child users on the same day, converting the activity duration, type and intensity into standard metabolic equivalent values ​​to generate activity level data.

[0068] Specifically, the system receives daily activity data from children uploaded via user terminals and collected by wearable devices. This data includes start and end timestamps for each activity period, activity type labeling, movement frequency data, heart rate change data, and environmental information. The system first preprocesses the received activity data, performing data type validation to ensure each data field conforms to preset specifications. It also performs logical consistency checks, verifying the reasonableness of the match between activity duration, movement frequency, and heart rate changes, and removing invalid data with ambiguous activity type labeling, logical conflicts between activity duration and intensity, or abnormal data formats. For valid data entries with missing information, the system combines the common characteristics of similar activities and the child's activity patterns at other times of the day to reasonably supplement the missing activity intensity information, ensuring the completeness and validity of the data.

[0069] After preprocessing, the system performs quantitative evaluation on the batch of data. It retrieves preset activity quantification conversion rules based on activity metabolic equivalent standards published by authoritative institutions. These rules establish a mapping relationship between activity type, intensity parameters, and standard metabolic equivalent values ​​according to the energy consumption characteristics of different activity types. This mapping relationship converts activity type and intensity parameters into corresponding standard metabolic equivalent baseline values. Based on this, the system performs weighted calculations on the standard metabolic equivalent baseline values ​​in conjunction with activity duration parameters. The calculation process fully considers the impact of the continuity of activities at different times on overall energy consumption, avoiding interference from data from a single time point on the overall evaluation results. The system integrates the quantification results corresponding to activities at each time period in a time series, and obtains the total standard metabolic equivalent value for the child user's activities for the day through cumulative calculation. This total value is then standardized to eliminate the influence of differences in basal metabolic rates among children of different age groups on the evaluation results, generating activity level data that characterizes the child's daily activity consumption level.

[0070] S23: Calculate the deviation of the personalized growth rate from the standard growth rate of the global benchmark knowledge base and the ratio of the activity level data to the benchmark activity level of the global benchmark knowledge base as dynamic adjustment factors. Input the dynamic adjustment factors into the preset personalized nutrition requirement prediction model, and perform weighted calculation of the standard nutrient reference intake and the dynamic adjustment factors to generate an individualized daily nutrient requirement vector.

[0071] Specifically, the system retrieves the generated personalized growth rate and the standard growth rate for the corresponding age group from the global benchmark knowledge base. The standard growth rate includes the baseline growth rate and growth rate variation range of children in that age group at different growth stages. The system compares and analyzes the two parameters through a preset deviation calculation logic, constructs a growth rate difference assessment function, and substitutes the stage characteristic values ​​and change slopes of the personalized and standard growth rates into the function to calculate the deviation of the personalized growth rate from the standard growth rate. Simultaneously, the system retrieves the generated activity level data and the benchmark activity level for the corresponding age group from the global benchmark knowledge base. The benchmark activity level includes the standard metabolic equivalent range and average consumption level of children in that age group's daily activities. The system calculates the activity level data and benchmark activity level using a preset ratio calculation logic. The calculation process follows a standardized numerical comparison procedure to ensure that the ratio result objectively reflects the difference between the two. The system integrates the calculated deviation and ratio through time-series correlation, and eliminates redundant information between the two types of parameters through a data fusion algorithm to form a unified dynamic adjustment factor, which includes growth dimension adjustment coefficients and activity dimension adjustment coefficients.

[0072] Furthermore, the system retrieves a pre-defined personalized nutrition requirement prediction model. This model, based on the physiological mechanisms of children's growth and development, constructs a correlation between dynamic adjustment factors and nutrient requirement adjustment amounts. The dynamic adjustment factors are input into this model, which performs feature extraction and hierarchical computation to generate corresponding nutrition requirement adjustment parameters. These parameters include personalized adjustment coefficients for different nutrients. The system retrieves standard nutrient reference intakes (NRAs) for the corresponding age group from the global benchmark knowledge base. These NRAs include basic intake standards for various macronutrients and micronutrients. The system performs a weighted calculation of the standard nutrient reference intakes and the adjustment parameters output by the model. The calculation process assigns corresponding weights according to the physiological functional characteristics of different nutrients, making targeted adjustments to the basic intake standards for various nutrients. The system integrates the weighted calculation results for each nutrient and sorts the results according to the nutrient type sorting rules defined in the global benchmark knowledge base, generating an individualized daily nutrient requirement vector. Each component of this vector corresponds to the individualized daily requirement of one nutrient.

[0073] In one embodiment, S3 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0074] S31: Perform feature extraction and classification processing on the food images uploaded by users, identify the food categories in the images through convolutional neural networks, estimate the physical quality of each type of food based on image features, and generate a set of recognition results containing food labels and quality pairs.

[0075] Specifically, the system receives food images uploaded by users via mobile terminals. These images are associated with a timestamp, dining scene annotations, and camera parameters. The system first preprocesses the food images, performing image noise reduction to eliminate ambient light interference and equipment noise, standardizing image size to conform to preset input specifications, and performing grayscale conversion and edge enhancement to improve the distinction between food and background areas. After preprocessing, the system extracts features from the images, including color, texture, and shape features. Color features are obtained through color histogram statistics, texture features are calculated using the gray-level co-occurrence matrix, and shape features are obtained through contour extraction and morphological analysis. The system inputs the extracted image features into a preset convolutional neural network. This network performs deep feature mining and dimensionality compression through multiple convolutional and pooling layers, and maps the extracted high-dimensional features into food category recognition vectors through fully connected layers. These vectors are then matched against a preset food feature database, outputting the food category recognition result and corresponding confidence level. For each identified food category, the system constructs a quality estimation model based on the food's pixel proportion, outline size, and associated tableware reference features in the image. The model then calculates the physical quality data for each food category. The system integrates the food category identification results with the corresponding quality data to generate a set of identification results containing the food's unique identifier, physical quality data, and identification confidence level.

[0076] S32: For each food identifier in the recognition result set, perform component mapping processing, query the global benchmark knowledge base to match its standard nutrient content data, and generate a nutrient vector corresponding to the recognized food.

[0077] Specifically, the system retrieves the recognition result set and extracts the food identification information. This food identification includes the food name, category code, and characteristic attribute tags, uniquely pointing to a specific food category. The system initiates a component mapping process, using the extracted food identification as search keywords to retrieve data from the global benchmark knowledge base for matching. The global benchmark knowledge base stores standard nutritional component data for various foods, labeled per unit mass, covering the specific components and content information of macronutrients and micronutrients. The system performs precise retrieval and matching through food category codes. If ambiguity occurs in the code matching, a secondary verification is performed using the food name and characteristic attribute tags to ensure the uniqueness and accuracy of the matching results. After matching, the system extracts the standard nutritional component data for the corresponding food and constructs a nutritional component vector based on this data. The dimension of the vector is consistent with the nutrient types defined in the global benchmark knowledge base, and each component of the vector corresponds to the content of the corresponding nutrient per unit mass of the food. The system performs validity verification on the generated nutritional component vector, checking whether the data of each component of the vector is complete and conforms to the basic logic of nutrient content. After passing the verification, a nutritional component vector corresponding one-to-one with each identified food category is generated.

[0078] S33: Based on the nutrient component vector and the corresponding food quality, a weighted sum is performed, and the contribution value of all meals is accumulated within the daily time window to generate a real-time dietary intake vector.

[0079] Specifically, the system retrieves nutrient component vectors and corresponding food quality data, multiplies each component in the nutrient component vector with the physical mass of the corresponding food, and obtains the actual intake value of various nutrients in a single food, thus quantifying the nutrient intake of a single food in a single meal. The system retrieves preset time window parameters for the day, which is defined as from 0:00 on the current day to the time the current meal image is uploaded, covering the breakfast, lunch, and dinner meals completed that day, as well as snack periods.

[0080] Furthermore, the system iterates through the recognition results and nutrient vectors corresponding to all uploaded meal images within the daily time window, and accumulates the intake values ​​of the same type of nutrient. During the accumulation process, the system records the intake contribution value corresponding to each meal in real time, ensuring that the accumulation result can be traced back to each meal. After the accumulation is completed, the system integrates the accumulated data, sorts the data according to the nutrient type sorting rules defined in the global benchmark knowledge base, and generates a real-time dietary intake vector. The dimension of this vector is completely consistent with the individualized daily nutrient requirement vector, and each component corresponds to the cumulative actual intake of one nutrient within the daily time window, providing direct data support for subsequent nutrient gap calculations.

[0081] In one embodiment, S4 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0082] S41: Perform gap analysis on the individualized daily nutrient requirement vector and the real-time dietary intake vector, compare each element to determine the difference between the immediate requirement and the intake of each nutrient, and generate the original nutrient gap.

[0083] Specifically, the system retrieves the generated individualized daily nutrient requirement vector and real-time dietary intake vector. The individualized daily nutrient requirement vector contains the individualized daily requirement value for each nutrient, the corresponding unique nutrient code, and the requirement priority identifier. The real-time dietary intake vector contains the cumulative actual intake value of each nutrient within the current time window, the contribution record of the corresponding meal, and the intake time distribution information. The system performs a dimensional consistency check on the two vectors, establishing a mapping relationship through nutrient codes and verifying whether the number of nutrient types and coding rules contained in the two are completely matched. If there is a dimensional mismatch, the system uses the nutrient classification standards of the global benchmark knowledge base to fill in missing nutrient entries with zero values ​​and merge duplicate entries for deduplication. After the verification is passed, the system constructs the original nutrient gap calculation model according to the logic of element-by-element comparison. The core formula is:

[0084]

[0085] in, This represents the initial nutrient deficit for the i-th nutrient. This represents the individualized daily requirement value for the i-th nutrient. This represents the cumulative actual intake of the i-th nutrient within the daily time window. The system encodes the corresponding nutrient in the two vectors. and The difference is calculated by substituting the values ​​into the formula. A positive result indicates an intake deficit for that nutrient, zero indicates adequate intake, and a negative result indicates excessive intake. The system integrates the difference results for all nutrients, labels the deficit status and corresponding deficit value for each nutrient, and generates a raw nutrient deficit data set containing nutrient codes, deficit values, deficit status, and demand priorities, thus completing the initial quantification of the nutrient deficit.

[0086] S42: Extract and parameterize the nutrient attribute rules in the global benchmark knowledge base, read the synergistic or antagonistic relationships between nutrients, the tolerable upper intake threshold, and the unit nutrient cost coefficient, and generate a set of rule parameters to describe nutrient interactions, safety limits, and economic constraints.

[0087] Specifically, the system retrieves nutrient attribute rule-related data from the global baseline knowledge base. This data is stored in structured tables, including nutrient association tables, safe intake threshold tables, and cost parameter tables. The system extracts nutrient attribute rules, retrieving three core types of data: first, synergistic or antagonistic interaction data between nutrients, including nutrient pairs with associated effects, effect type identifiers, and descriptions of the degree of influence; second, tolerable upper intake level threshold data, including upper intake limit standards for each nutrient, categorized by age group; and third, unit nutrient cost coefficient data, including unit nutrient content cost parameters for basic supplementary ingredients for various nutrients. The system then performs parameterization processing on the extracted three types of data, converting the descriptions of the degree of influence into standardized numerical coefficients to construct a nutrient interaction parameter matrix. The core expression is:

[0088]

[0089] in, This represents the interaction parameter between the i-th nutrient and the j-th nutrient. =1 indicates synergy. =-1 indicates an antagonistic effect. =0 indicates no direct effect. Simultaneously, the system uses structured encoding of the tolerable upper intake level threshold to obtain the safety threshold parameter. Cost parameters are obtained by standardizing the unit nutrient cost coefficient. The system integrates the parameterized interaction parameter matrix, safety threshold parameters, and cost parameters to generate a set of rule parameters, which is indexed according to nutrient codes.

[0090] S43: Multi-objective fusion modeling is performed based on the original nutrient gap and the set of rules and parameters. The original nutrient gap is used as the baseline objective. At the same time, the synergistic and antagonistic relationship data in the set of rules and parameters are introduced to adjust the supplementation ratio of various nutrients. The safety and cost dual constraints of the tolerable maximum intake threshold and the unit nutrient cost coefficient extracted from the set of rules and parameters are superimposed to construct a nutritional complementarity optimization model under multiple constraints.

[0091] Specifically, the system constructs a gap-filling rate evaluation function with the degree of filling of the original nutrient gap as the benchmark target. ,in This represents the planned intake of the i-th nutrient. This represents the total number of nutrient types, and the function is used to quantify the degree of gap filling. Based on this, the system extracts the interaction parameter matrix from the set of rule parameters to construct a nutrient co-optimization function. This function positively correlates and regulates the intake of nutrients with synergistic effects, while negatively constraining and regulating the intake of nutrients with antagonistic effects. Simultaneously, the system extracts the safety threshold parameter $$U_i$$ and cost parameter $$ from the rule parameter set. Construct dual constraints, with the safety constraint being... Ensure that the total intake after supplementation does not exceed the tolerable upper limit; cost constraints are:

[0092]

[0093] in, This represents the preset upper limit of cost per supplementation. The system integrates the baseline objective function, the collaborative optimization function, and dual constraints. Through variable correlation mapping, it establishes logical connections between modules, constructing a nutrient complementarity optimization model under multiple constraints. The decision variables of the model are the supplementation amounts of various nutrients. .

[0094] S44: Iteratively solve the nutritional complementarity optimization model. Combine the preset sequential quadratic programming algorithm to iteratively adjust and evaluate the candidate nutritional supplementation schemes under the dual constraints of safety and cost. When the rate of change of the overall optimization objective function value corresponding to the candidate nutritional supplementation scheme is less than the preset convergence threshold during the iteration process, it is determined that it has converged to the optimal interval, and the current candidate scheme is output as the theoretically optimal nutritional supplementation vector.

[0095] Specifically, the system retrieves the constructed nutrient complementarity optimization model and clarifies the decision variables in the model. Objective function (in , The specific expressions for the weighting coefficients of the gap-filling objective and the collaborative optimization objective, respectively, and the constraints, are used to determine the variable boundaries of the supplementation amounts for various nutrients. The system calls a preset sequential quadratic programming algorithm to convert the model parameters into an algorithm-recognizable input format and iteratively solves the problem. The core expression of the sequential quadratic programming algorithm is:

[0096]

[0097]

[0098]

[0099] in, This is the nutrient replenishment vector at the k-th iteration. This is the search direction vector for the current iteration point. Let be the local curvature matrix of the objective function at the k-th iteration. For the objective function in The gradient vector at that point, Let i be the constraint function for the i-th equality. Let j be the constraint function for the j-th inequality. These are the set indices for equality constraints and inequality constraints, respectively. Based on this algorithm, the system calculates the current search direction vector and the gradient of the objective function in each iteration, adjusts the candidate nutrient supplementation schemes, and simultaneously verifies whether the schemes meet both safety and cost constraints, calculating the overall optimization objective function value corresponding to each scheme. The system constructs convergence criteria:

[0100]

[0101] in, The preset convergence threshold, These are the objective function values ​​for the (k+1)th and kth iterations, respectively. When the convergence criterion is met, the system determines that the iteration process has converged to the optimal interval and stops iterating. The system extracts candidate nutrient supplementation schemes corresponding to the current iteration, performs a final verification on the amount of each nutrient supplemented in the scheme, and ensures that it meets the original nutrient gap filling requirements and all constraints. After the verification is passed, the scheme is output as the theoretically optimal nutrient supplementation vector.

[0102] In one embodiment, S5 of the personalized nutrition package planning method based on children's health management provided by the present invention specifically includes the following steps:

[0103] S51: Initialize and characterize the pre-stored nutrient component library to ensure that each basic nutrient ingredient is associated with a component feature vector defined based on the content of each nutrient in the basic nutrient ingredient, and generate a standardized component feature matrix.

[0104] Specifically, the system retrieves a pre-stored nutrient component library, which contains information on various basic nutritional ingredients, including ingredient name, unique code, detailed nutrient composition, purity level, storage characteristics, and application-related parameters such as dosage form characteristics (such as solubility, dispersibility, and stability) and adjustable concentration range suitable for children's nutritional packs. The system initiates an initialization process, performing a dual check on the completeness and application suitability of the ingredient information in the nutrient component library. It verifies whether the detailed nutrient composition of each ingredient is complete and the code is unique, while also checking whether the dosage form characteristics meet the production and usage requirements of children's nutritional packs. Ingredient entries with missing information or incompatible dosage forms are marked and removed, and duplicate codes are merged.

[0105] After initialization, the system performs feature processing, extracting the content data of various nutrients from the nutrient composition details of each basic nutritional ingredient, and constructing a component feature vector by combining the adjustable concentration range. ,in Let m represent the component feature vector of the m-th basic nutrient raw material. This represents the unit mass content of the i-th nutrient in the m-th raw material. The system represents the total number of nutrient types, consistent with the theoretically optimal nutrient supplementation vector dimension. It integrates the component feature vectors of all basic nutrient raw materials according to their coding order, while also associating the dosage form characteristic parameters of each raw material to generate a standardized component feature matrix. The matrix consists of rows corresponding to basic nutritional ingredients, columns corresponding to nutrient types, and additional columns indicating ingredient dosage form compatibility information. The system validates the component feature matrix to ensure there are no zero vector rows, no outliers, and complete dosage form compatibility information. After successful validation, the initialization and feature processing of the nutritional component library are completed.

[0106] S52: Match and fit the theoretically optimal nutrient supplementation vector with the component feature matrix, and transform it into a linear programming problem with the goal of minimizing the ratio error and the constraint of non-negativity of raw materials. This will generate a raw material weight ratio scheme that meets the requirements of the theoretically optimal nutrient supplementation vector and has the best cost.

[0107] Specifically, the system retrieves the theoretically optimal nutrient supplementation vector. With component feature matrix The system initiates the matching and fitting process, simultaneously retrieving actual production process constraints of the children's nutritional supplement packs (such as the total weight range per pack and requirements for raw material mixing uniformity). The system first performs a dimensionality consistency check to ensure that the number of columns in the component feature matrix perfectly matches the dimensions of the theoretically optimal nutritional supplement vector. If a mismatch exists, the system completes or adjusts the column dimensions of the component feature matrix based on the nutrient classification standards of the global benchmark knowledge base. After successful verification, the system transforms the matching and fitting problem into a linear programming problem that balances nutritional goals and production applications, constructing the objective function:

[0108]

[0109] in, Indicates the error in the mixing ratio. For the decision variable vector, This represents the weight of the m-th basic nutrient ingredient. Simultaneously, the system sets multiple constraints: one is... First, ensure that the weight of the raw materials is non-negative; second, These are the minimum and maximum weight thresholds for a single nutrient pack, respectively, to adapt to actual packaging specifications; thirdly, based on the dosage form characteristic parameters in the component feature matrix, raw material mixing compatibility constraints are set to ensure that different raw materials meet stability requirements after mixing.

[0110] Furthermore, the system invokes a preset linear programming algorithm, substituting the objective function and multiple constraints into the algorithm for solution. During the solution process, the decision variables are iteratively optimized. By taking the appropriate value, the mixing ratio error can be continuously reduced. Simultaneously, it ensures that the production process and mixing compatibility requirements are met until the error value is less than the preset error threshold. After the solution is completed, the system extracts the optimal values ​​of the decision variables and generates a raw material weight ratio scheme that includes the codes of each basic nutrient raw material, corresponding weights, dosage form adaptation instructions, and mixing compatibility prompts. This scheme not only meets the nutritional objectives of the theoretically optimal nutrient supplementation vector but also meets the process requirements and cost optimization requirements of actual nutrient pack production.

[0111] S53: Format and generate instructions for the raw material weight ratio scheme, converting it into standardized text and lists containing raw material names, precise mass, mixing order and preparation instructions, and generating an executable nutrition pack formula.

[0112] Specifically, the system performs dual verification of the raw material weight ratio scheme for effectiveness and production adaptability. It checks whether the weight of each raw material is within the preset reasonable range, whether the raw material combination can fully cover the required nutrients, and whether the ratio scheme meets the feeding requirements of the production equipment. Any abnormal ratio data is corrected. After successful verification, the system converts the raw material codes in the raw material weight ratio scheme into their corresponding raw material names and standardizes the weight data according to the required production accuracy (e.g., retaining the decimal places suitable for the weighing equipment). Following a dual-adaptation format for industrial production and user use, the system integrates raw material names and precise quality data, combining dosage form characteristics and production process constraints in the component feature matrix to generate production execution instructions. These instructions include key parameters for each production stage, such as the raw material feeding sequence, mixing speed and time, and drying temperature and duration. Simultaneously, it retrieves child-appropriate usage parameters stored in the global baseline knowledge base to supplement and generate user instructions, including preparation temperature, solvent usage, stirring requirements, consumption time (e.g., how long after a meal), and single-dose dosage for children of different age groups.

[0113] Furthermore, the system integrates production execution instructions and user manuals into standardized text and lists. The text details the core production requirements and usage precautions of the formula, while the lists clearly present the key parameters of each raw material, the sequence of production steps, and key points of operation. The system performs format verification and practicality review on the standardized text and lists to ensure accurate expression, clear logic, and unambiguity, and that they conform to the instruction receiving standards of the production line and the user's understanding and usage habits. After verification, an executable nutritional package formula is generated. This formula can be directly used to guide the actual production and preparation of nutritional packages, while providing users with clear usage guidelines, achieving a complete implementation from theoretical solutions to actual products.

[0114] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0115] Based on the same inventive concept, this application also provides a device for planning personalized nutrition packages based on children's health management to implement the aforementioned method for planning personalized nutrition packages based on children's health management. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the device for planning personalized nutrition packages based on children's health management provided below can be found in the limitations of the method for planning personalized nutrition packages based on children's health management described above, and will not be repeated here.

[0116] Preferably, such as Figure 2 As shown, the present invention provides a personalized nutrition package planning device 600 based on children's health management, which is configured with the following modules:

[0117] The nutrition profile construction module 610 is used to integrate the identity information, continuously recorded anthropometric data and biochemical index data in the original data of the collected children users into an individualized nutrition profile and to build a global benchmark knowledge base containing standard food nutrient component vectors and nutrient attribute rules based on a pre-set nutrition standard database.

[0118] The daily nutrient requirement calculation module 620 is used to fit the personalized growth rate of children based on the time-series growth data in the individualized nutrition profile, and to calculate and generate an individualized daily nutrient requirement vector covering multiple nutrients by combining the obtained activity level data of the children on that day with the standard reference value.

[0119] The dietary intake calculation module 630 is used to identify and process food images uploaded by users through the terminal, identify food types and estimate the quality of various identified foods, query the global benchmark knowledge base to match nutritional component data and accumulate calculations to generate a real-time dietary intake vector.

[0120] The nutrition supplementation optimization module 640 is used to calculate the original nutritional gap based on the individualized daily nutritional requirement vector and the real-time dietary intake vector, and to construct a nutritional complementarity optimization model based on the nutrient interaction rules and safety constraints in the global benchmark knowledge base, and calculate the optimized nutritional supplementation plan, and output the theoretically optimal nutritional supplementation vector.

[0121] The nutrition pack formula generation module 650 is used to match and fit the theoretical optimal nutrition supplementation vector with the pre-stored nutrition component library, solve for the weight ratio of specific raw materials that meet the nutritional goals defined by the theoretical optimal nutrition supplementation vector, and generate the nutrition pack formula. The nutrition pack formula is used to indicate the ratio of nutrients to supplement the nutritional gaps required by child users on that day.

[0122] Preferably, the daily nutrient requirement calculation module 620 provided in this application is configured with the following units:

[0123] The growth rate fitting unit is used to perform trend analysis on continuously recorded anthropometric data in individualized nutrition records, apply time series models to fit growth curves and calculate rate changes to generate personalized growth rates.

[0124] The activity level quantification unit is used to quantify and evaluate the activity data of the child users on the same day, converting the activity duration, type and intensity into standard metabolic equivalent values ​​to generate activity level data;

[0125] The daily nutrient requirement prediction unit is used to calculate the deviation of the personalized growth rate from the standard growth rate of the global benchmark knowledge base and the ratio of the activity level data to the benchmark activity level of the global benchmark knowledge base as dynamic adjustment factors. The dynamic adjustment factors are input into the preset personalized nutrient requirement prediction model, and the standard nutrient reference intake and the dynamic adjustment factors are weighted and calculated to generate an individualized daily nutrient requirement vector.

[0126] Preferably, the dietary intake calculation module 630 provided in this application is configured with the following units:

[0127] The food image recognition unit is used to extract features and classify food images uploaded by users. It identifies the food categories in the images through convolutional neural networks and estimates the physical quality of each type of food based on image features, generating a set of recognition results containing food labels and quality pairs.

[0128] The food composition mapping unit is used to perform composition mapping processing on each food identifier in the recognition result set, query the global benchmark knowledge base to match its standard nutrient content data, and generate a nutrient vector corresponding to the identified food.

[0129] The dietary intake accumulation unit is used to perform weighted summation based on the nutrient component vector and the corresponding food quality, and to accumulate the contribution value of all meals within the daily time window to generate a real-time dietary intake vector.

[0130] Preferably, the nutrition supplementation optimization module 640 provided in this application is configured with the following units:

[0131] The nutrient gap analysis unit is used to perform gap analysis on the individualized daily nutrient requirement vector and the real-time dietary intake vector, compare each element to determine the difference between the immediate requirement and the intake of each nutrient, and generate the original nutrient gap.

[0132] The rule parameter extraction unit is used to extract and parameterize the nutrient attribute rules in the global benchmark knowledge base. It reads the synergistic or antagonistic relationship between nutrients, the tolerable upper intake threshold, and the unit nutrient cost coefficient, and generates a set of rule parameters to describe nutrient interactions, safety limits, and economic constraints.

[0133] The nutrition optimization modeling unit is used to perform multi-objective fusion modeling based on the original nutritional gap and the set of rules and parameters. The original nutritional gap is used as the baseline objective. At the same time, the synergistic and antagonistic relationship data in the set of rules and parameters are introduced to adjust the supplementation ratio of various nutrients. The safety and cost dual constraints are superimposed by the tolerable maximum intake threshold and the unit nutritional cost coefficient extracted from the set of rules and parameters, and a nutritional complementarity optimization model under multiple constraints is constructed.

[0134] The optimization model solving unit is used to iteratively solve the nutritional complementarity optimization model. Combined with the preset sequential quadratic programming algorithm, it iteratively adjusts and evaluates the candidate nutritional supplementation schemes under the dual constraints of safety and cost. When the rate of change of the overall optimization objective function value corresponding to the candidate nutritional supplementation scheme is less than the preset convergence threshold during the iteration process, it is determined that it has converged to the optimal interval, and the current candidate scheme is output as the theoretically optimal nutritional supplementation vector.

[0135] Preferably, the nutrition pack formula generation module 650 provided in this application is configured with the following units:

[0136] The nutrient component characterization unit is used to initialize and characterize the pre-stored nutrient component library, ensuring that each basic nutrient raw material is associated with a component feature vector defined based on the content of each nutrient in the basic nutrient raw material, and generating a standardized component feature matrix.

[0137] The raw material ratio optimization solution unit is used to match and fit the theoretical optimal nutrient supplementation vector and the component feature matrix, and transform it into a linear programming problem with the goal of minimizing the ratio error and the constraint of non-negativity of raw materials, to generate a raw material weight ratio scheme that meets the requirements of the theoretical optimal nutrient supplementation vector and has the best cost.

[0138] The nutrition pack formula generation unit is used to format and generate instructions for the raw material weight ratio scheme, converting it into standardized text and lists containing raw material names, precise mass, mixing order and preparation instructions, and generating an executable nutrition pack formula.

[0139] In one embodiment, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described personalized nutrition package planning method based on children's health management.

[0140] In one embodiment, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described personalized nutrition package planning method based on children's health management.

[0141] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0142] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0143] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A personalized nutrition package planning method based on children's health management, characterized in that, Includes the following steps: S1: Integrate the identity information, continuously recorded anthropometric data and biochemical index data from the collected raw data of child users into individualized nutrition profiles, and build a global benchmark knowledge base containing standard food nutrient component vectors and nutrient attribute rules based on a pre-set nutrition standard database. S2: Based on the time-series growth data in the individualized nutrition profile, fit the child user's personalized growth rate, and combine the obtained child user's daily activity level data with standard reference values ​​to calculate and generate an individualized daily nutrient requirement vector covering multiple nutrients. S3: Recognize and process the food images uploaded by the user through the terminal, identify the types of food and estimate the quality of each identified food, query the global benchmark knowledge base to match the nutritional component data and accumulate and calculate to generate a real-time dietary intake vector. S4: Calculate the original nutritional gap based on the individualized daily nutrient requirement vector and the real-time dietary intake vector, and construct a nutritional complementarity optimization model based on the nutrient interaction rules and safety constraints in the global benchmark knowledge base, calculate the optimized nutritional supplementation plan, and output the theoretically optimal nutritional supplementation vector. S5: Match and fit the theoretically optimal nutritional supplementation vector with the pre-stored nutritional component library to solve for the weight ratio of specific raw materials that meet the nutritional goals defined by the theoretically optimal nutritional supplementation vector, and generate a nutritional pack formula. The nutritional pack formula is used to indicate the ratio of nutrients that supplement the nutritional gaps required by child users on that day.

2. The method according to claim 1, characterized in that, S2 includes: S21: Perform trend analysis on the continuously recorded anthropometric data in the individualized nutrition record, apply a time series model to fit the growth curve and calculate the rate change to generate the personalized growth rate. S22: Quantitatively evaluate and process the acquired activity data of the child user for the day, converting the activity duration, type and intensity into standard metabolic equivalent values ​​to generate activity level data; S23: Calculate the deviation of the personalized growth rate from the standard growth rate of the global benchmark knowledge base and the ratio of the activity level data to the benchmark activity level of the global benchmark knowledge base as dynamic adjustment factors. Input the dynamic adjustment factors into a preset personalized nutrition requirement prediction model, and perform weighted calculation of the standard nutrient reference intake and the dynamic adjustment factors to generate an individualized daily nutrient requirement vector.

3. The method according to claim 1, characterized in that, S3 includes: S31: Perform feature extraction and classification processing on the food images uploaded by users, identify the food categories in the images through convolutional neural networks, estimate the physical quality of each type of food based on image features, and generate a set of recognition results containing food labels and quality pairs; S32: Each food identifier in the recognition result set is subjected to component mapping processing, and the nutritional component content data of its standard is matched by querying the global benchmark knowledge base to generate a nutritional component vector corresponding to the identified food. S33: Based on the nutrient component vector and the corresponding food quality, perform a weighted summation and accumulate the contribution values ​​of all meals within the daily time window to generate a real-time dietary intake vector.

4. The method according to claim 1, characterized in that, S4 includes: S41: Perform gap analysis on the individualized daily nutrient requirement vector and the real-time dietary intake vector, compare each element to determine the difference between the immediate requirement and the intake of each nutrient, and generate the original nutrient gap. S42: Extract and parameterize the nutrient attribute rules in the global benchmark knowledge base, read the synergistic or antagonistic relationship between nutrients, the tolerable maximum intake threshold, and the unit nutrient cost coefficient, and generate a set of rule parameters to describe nutrient interactions, safety limits, and economic constraints. S43: Based on the original nutritional gap and the set of rules parameters, perform multi-objective fusion modeling. With the original nutritional gap as the baseline objective, introduce the synergistic and antagonistic relationship data in the set of rules parameters to adjust the supplementation ratio of various nutrients. Superimpose the dual constraints of safety and cost formed by the tolerable maximum intake threshold and the unit nutrient cost coefficient extracted from the set of rules parameters to construct a nutritional complementarity optimization model under multiple constraints. S44: The nutritional complementarity optimization model is iteratively solved. The candidate nutritional supplementation schemes are iteratively adjusted and evaluated under the dual constraints of safety and cost by combining the preset sequential quadratic programming algorithm. When the rate of change of the overall optimization objective function value corresponding to the candidate nutritional supplementation scheme is less than the preset convergence threshold during the iteration process, it is determined that the scheme has converged to the optimal interval, and the current candidate scheme is output as the theoretically optimal nutritional supplementation vector.

5. The method according to claim 4, characterized in that, The expression for the sequential quadratic programming algorithm is: Where x represents the planned supplementation amount of all nutrients to be optimized. The vector formed, i.e. ,in This represents the amount of the i-th nutrient to be supplemented. Let be the nutrient replenishment vector at the k-th iteration. At the current iteration point The search direction vector is to be found at the location. To describe the local curvature of the objective function at the k-th iteration, The objective function f(x) of the original problem is located at point... The gradient vector at that point, Let i be the i-th equality constraint function, and let its set index be . , Let I be the j-th inequality constraint function, and let I be its set index.

6. The method according to any one of claims 1-5, characterized in that, S5 includes: S51: Initialize and characterize the pre-stored nutrient component library to ensure that each basic nutrient raw material is associated with a component feature vector defined based on the content of each nutrient in the basic nutrient raw material, and generate a standardized component feature matrix. S52: Match and fit the theoretical optimal nutrient supplementation vector with the component feature matrix, and transform it into a linear programming problem with the goal of minimizing the ratio error and the constraint of non-negativity of raw materials, to generate a raw material weight ratio scheme that meets the requirements of the theoretical optimal nutrient supplementation vector and has the best cost. S53: The raw material weight ratio scheme is formatted and generated by instruction generation, and converted into standardized text and list containing raw material names, precise mass, mixing order and preparation instructions, to generate an executable nutrition pack formula.

7. A personalized nutrition package planning device based on children's health management, characterized in that, The device includes: The nutrition profile building module is used to integrate the identity information, continuously recorded anthropometric data and biochemical index data in the original data of collected children users into an individualized nutrition profile and build a global benchmark knowledge base containing standard food nutrient component vectors and nutrient attribute rules based on a pre-set nutrition standard database. The daily nutrient requirement calculation module is used to fit the child user's personalized growth rate based on the time-series growth data in the individualized nutrition profile, and to calculate and generate an individualized daily nutrient requirement vector covering multiple nutrients by combining the obtained activity level data of the child user on that day with the standard reference value. The dietary intake calculation module is used to identify and process food images uploaded by users through the terminal, identify food types and estimate the quality of each identified food, query the global benchmark knowledge base to match nutritional component data and accumulate calculations to generate a real-time dietary intake vector. The nutrition supplementation optimization module is used to calculate the original nutritional gap based on the individualized daily nutritional requirement vector and the real-time dietary intake vector, and to construct a nutritional complementarity optimization model based on the nutrient interaction rules and safety constraints in the global benchmark knowledge base and calculate the optimized nutritional supplementation plan, and output the theoretically optimal nutritional supplementation vector. The nutrition pack formula generation module is used to match and fit the theoretically optimal nutrition supplementation vector with a pre-stored nutrition component library, solve for the weight ratio of specific raw materials that meet the nutritional goals defined by the theoretically optimal nutrition supplementation vector, and generate a nutrition pack formula. The nutrition pack formula is used to indicate the ratio of nutrients that supplement the nutritional gaps required by child users on that day.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

Images