Ai nutrition screening detector and detection method

Abstract

The invention provides an Ai nutrition screening detector and a detection method, when the Ai nutrition screening detector is applied, an automatic screening process can be carried out: step 1, a user stands on a detection platform in a barefoot manner, an electrode automatically identifies foot contact, and BIA scanning is synchronously completed within 60 seconds; the millimeter-wave radar captures a complete respiratory cycle; the optical system generates a 3D volume mesh model; 2, the detector uploads detection data to an edge server; and step 3, the edge computing node operates the MFN nutrition risk detection model to generate a nutrition risk report. Intelligent nutrition detection analysis is achieved, user nutrition risk scores and intervention suggestions are output, multi-modal nutrition index analysis and fusion prediction are achieved, multi-modal nutrition screening detection is achieved, the application range is large, the method is suitable for detection of different user crowds, AI technical support is utilized, the defects of manual calculation and analysis and single-computer calculation and analysis are overcome, and the user experience is improved. The AI technology is fused into the nutrition screening detection technology, and the intelligent detection effect is achieved.

Description

Technical Field

[0001] This invention relates to the field of health management technology, and in particular to an AI nutrition screening and testing instrument and method, electronic device and computer-readable storage medium. Background Technology

[0002] Nutritional screening and testing are examination items that are gradually gaining attention, aiming to focus on dietary balance and disease screening.

[0003] Existing nutritional screening and testing technologies / equipment solutions, and commonly used tools in clinical and community settings, mainly include:

[0004] 1. Standardized screening scale

[0005] For example, MUST (Meditation Screening Tool): rapidly stratifies risk based on BMI, weight loss, and disease status score. NRS-2002 (Nutritional Risk Screening): combines disease severity with nutritional status score (European clinical recommendation). SGA (Subjective Comprehensive Assessment): qualitatively assesses nutritional status through medical history and physical examination.

[0006] 2. Electronic screening platform

[0007] Integrated screening system: Hospital information system (HIS / EHR integrated module), automatic BMI calculation, screening score (e.g., embedded NRS-2002). The system can provide early warnings for high-risk patients (e.g., automatically triggering a nutrition consultation upon hospitalization).

[0008] 3. Body composition analysis equipment

[0009] Bioelectrical impedance analysis (BIA), such as portable body composition analyzers (e.g., InBody, Tanita), estimates body fluid, muscle mass, and fat mass by measuring differences in the electrical resistance of biological tissues.

[0010] 4. Biochemical and metabolic detection

[0011] Laboratory analysis included blood parameters: serum albumin (<3.5g / dL indicates risk, long half-life); prealbumin (short half-life, reflects recent changes); lymphocyte count (assessing immunonutritional status); vitamin D (25-OH-D); and trace elements (Zn / Fe).

[0012] etc.

[0013] Although the above methods can perform nutritional screening in different dimensions, they are almost all single-modal nutritional indicator detection and cannot achieve multimodal nutritional screening detection, thus limiting their application scope.

[0014] Furthermore, the aforementioned nutritional screening tests mainly rely on manual and computer calculation and analysis, lacking AI technology analysis and failing to integrate AI technology into nutritional screening and testing technology, thus lacking intelligent testing effects.

[0015] Therefore, existing nutritional screening and testing methods are limited in scope and lack intelligence, and are not supported by AI technology. Summary of the Invention

[0016] To address the technical problems existing in the prior art, the present invention provides the following technical solution:

[0017] On one hand, an AI nutrition screening and testing device is provided, the AI ​​nutrition screening and testing device including a testing platform 1 and a support 6 disposed on the testing platform 1, the AI ​​nutrition screening and testing device further including:

[0018] Foot BIA electrode module 2 is used to acquire bioelectrical impedance data of the user's foot and feed it back to the industrial control computer 5;

[0019] 3D volumetric mesh optical system 3 is used to acquire volumetric mesh point cloud data of user-specified parts and feed it back to industrial control computer 5;

[0020] The millimeter-wave radar respiratory detection unit 4 is used to scan and detect the periodic displacement change signal generated by the user's chest cavity during breathing and feed it back to the industrial control computer 5.

[0021] The industrial control computer 5 is used to evaluate and calculate the user's BIA body composition data, respiratory rate variability index and body grid volume based on the bioelectrical impedance data, volume grid point cloud data and displacement periodic change signal generated by the thoracic cavity, and upload them to the edge server through the industrial gateway.

[0022] An industrial gateway is used to provide data communication services between the AI ​​nutrition screening and testing instrument and the edge server;

[0023] Power supply, used for supplying power;

[0024] The foot BIA electrode module 2, the 3D volumetric mesh optical system 3, the millimeter-wave radar respiratory detection unit 4, the industrial gateway, and the power supply are all electrically connected to the industrial control computer 5.

[0025] The industrial control computer 5 communicates with the edge server through the industrial gateway.

[0026] Preferably, the BIA body composition data includes BIA muscle mass fraction, the evaluation formula of which is as follows:

[0027] ,

[0028] in,

[0029] SM represents muscle mass fraction;

[0030] H: Height (unit: centimeters), reflecting the longitudinal dimension of the human body;

[0031] R: Foot-hand bioelectrical impedance (unit: ohm), characterizing the electrical conductivity of body fluids;

[0032] S: Gender coefficient (male = 1, female = 0), used for gender difference correction;

[0033] A: Age (unit: years), considering the age-related degenerative effects of muscle.

[0034] Preferably, the calculation formula for evaluating the respiratory rate variability index is as follows:

[0035] ,

[0036] in,

[0037] SDNN represents the respiratory rate variability index;

[0038] RR i : Duration of the i-th respiratory cycle (ms);

[0039] : The average respiratory interval over N cycles;

[0040] N: Number of sampling periods (fixed to 10);

[0041] ∑: Sum of the squares of the time deviations over consecutive respiratory cycles;

[0042] √: Standard deviation calculation, quantifying the degree of fluctuation in breathing rhythm.

[0043] Preferably, the formula for evaluating the volumetric mesh volume is as follows:

[0044] ,

[0045] in,

[0046] V represents the volume of the user's body grid;

[0047] v0, v1, v2, v3: The coordinates of the tetrahedral vertices of the user's body, derived from 3D camera point cloud data;

[0048] M: The set of all tetrahedra contained in the volume mesh model;

[0049] ||: Take the absolute value to ensure that the volume is non-negative.

[0050] On the other hand, a detection method for an AI nutrition screening device is provided, the detection method comprising the following steps:

[0051] The system initializes, establishes data communication between the AI ​​nutrition screening and testing instrument and the edge server, and constructs the user's nutrition testing report;

[0052] The edge server issues a sampling command and forwards it to the industrial control computer 5 of the Ai nutrition screening and testing instrument through the industrial gateway;

[0053] The industrial control computer 5 activates the millimeter-wave radar respiratory detection unit 4 and determines whether it has received the user's respiratory rate variability data:

[0054] If no signal is received, it indicates that there is no user or the user's posture is incorrect, and a detection alarm signal is sent to the edge server through the industrial gateway.

[0055] If received, the foot BIA electrode module 2, 3D volumetric mesh optical system 3 and millimeter-wave radar respiratory basket detection unit 4 will start sampling according to the preset sampling frequency;

[0056] The industrial control computer 5 acquires the sampling data, and based on the bioelectrical impedance data, volume mesh point cloud data and displacement periodic change signal generated by the thoracic cavity obtained from the sampling, it evaluates and calculates the user's BIA body composition data, respiratory rate variability index and volume mesh volume, and uploads them to the edge server through the industrial gateway.

[0057] The edge server receives and records the user's BIA body composition data, respiratory rate variability index, and body grid volume into the nutrition test report;

[0058] The edge server simultaneously inputs the user's BIA muscle mass score, respiratory rate variability index, and body grid volume into the preset MFN nutritional risk detection model. The MFN nutritional risk detection model identifies and outputs the corresponding nutritional risk score and intervention suggestions, and records the nutritional risk score and intervention suggestions in the nutritional detection report.

[0059] Preferably, the detection method further includes the following steps:

[0060] The user's nutrition test report is synchronously pushed to the APP / PC management terminal through the edge server.

[0061] Preferably, the method for generating the MFN nutritional risk detection model includes:

[0062] Collect and preprocess historical nutritional consultation big data from several users. The historical nutritional consultation big data includes users' BIA body composition data, respiratory rate variability index and body grid volume, as well as nutritional risk scores determined by quantitative assessment forms and intervention recommendations issued by nutritionists during clinical consultations.

[0063] Feature engineering was performed on the historical nutritional visit big data to extract corresponding BIA body composition features, respiratory rate variability features and volume grid volume features, and feature fusion was performed to generate historical nutritional visit features for each user.

[0064] The historical nutritional consultation characteristics of each user, as well as the corresponding nutritional risk scores and intervention suggestions, are statistically analyzed to generate a feature set consisting of several {historical nutritional consultation characteristics, nutritional risk scores, and intervention suggestions}, which are then divided into a training set and a validation set according to a preset ratio.

[0065] The training set is input into the preset MFN model to perform feature classification learning and generate an initial MFN nutrition risk detection model.

[0066] The validation set is input into the MFN nutrition risk detection model to verify its predictive and identification performance.

[0067] If the verification is successful, the MFN nutrition risk detection model is deployed and applied to the edge server;

[0068] Conversely, repeat the above steps to reconstruct the MFN nutritional risk detection model.

[0069] On the other hand, an electronic device is provided, comprising: a processor; and a memory storing computer-readable instructions, which, when executed by the processor, implement any of the detection methods described above.

[0070] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored therein, the at least one instruction being loaded and executed by a processor to implement any of the methods described above in the detection method.

[0071] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following:

[0072] When the AI-powered nutritional screening instrument of this invention is applied, it can automate the screening process as follows: Step 1: The user stands barefoot on the testing platform, the electrodes automatically identify foot contact, and a BIA scan is completed synchronously within 60 seconds; millimeter-wave radar captures 10 complete respiratory cycles; the optical system generates a 3D volumetric mesh model; Step 2: The instrument uploads the test data to the edge server; Step 3: The edge computing node runs the MFN nutritional risk detection model and generates a nutritional risk report. It can achieve intelligent nutritional detection and analysis, outputting a user's nutritional risk score and intervention suggestions, realizing multimodal nutritional index analysis and fusion prediction, achieving multimodal nutritional screening detection. It has a wide range of applications, suitable for different user groups, and utilizes AI technology to overcome the shortcomings of manual and single-computer calculation and analysis, integrating AI technology into nutritional screening detection technology to achieve intelligent detection results. Attached Figure Description

[0073] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0074] Figure 1 This is a schematic diagram of the application structure of an AI nutrition screening and testing instrument provided in an embodiment of the present invention;

[0075] Figure 2 This is a schematic diagram of the control system structure of an AI nutrition screening and testing instrument provided in an embodiment of the present invention;

[0076] Figure 3 This is a schematic diagram illustrating the application components of an AI nutrition screening and detection system provided in an embodiment of the present invention;

[0077] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0078] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0079] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0080] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0081] In this embodiment of the invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0082] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0083] This invention provides an AI nutrition screening instrument and a detection method. The method can be implemented by an electronic device, which can be a terminal or a server.

[0084] like Figure 1 The diagram shown illustrates the application principle of the AI ​​nutrition screening device. The device includes a detection platform 1 and a support 6 mounted on the platform 1. Similar to existing height measurement devices, the platform has a foot BIA electrode module 2 on its upper surface. A support also supports a 3D volumetric mesh optical system 3, a millimeter-wave radar respiratory basket detection unit 4, and an industrial control computer 5. The specific installation height and position of these components are not limited in this embodiment. Other supporting facilities, such as power supplies, displays, alarms, and gateways, can be installed independently.

[0085] When applied, the automated screening process is as follows: Step 1: The user stands barefoot on the testing platform, the electrodes automatically identify foot contact, and the BIA scan is completed synchronously within 60 seconds; the millimeter-wave radar captures 10 complete respiratory cycles; the optical system generates a 3D volumetric mesh model; Step 2: The detector uploads the test data to the edge server; Step 3: The edge computing nodes (specifically deployed edge servers, such as one in a unit / department, etc., and each node can form a nutrition testing topology network to achieve large-area / multi-region nutrition screening, such as the health management department ultimately achieving massive user nutrition screening tests at each node in the region) run the MFN nutrition risk detection model and generate a nutrition risk report.

[0086] The display screen can show user detection data and reports issued from the backend (the current user's detection report can be synchronously sent to the display screen by the edge server). When no one is detected standing on the platform for measurement, an alarm or corresponding prompt can be issued. The application system and AI detection principle of this invention will be further described below.

[0087] like Figure 2 As shown, the Ai nutrition screening and testing instrument also includes the following testing system:

[0088] Foot BIA electrode module 2 is used to acquire bioelectrical impedance data of the user's feet and feed it back to the industrial control computer 5; for example, it adopts a 4-pole 8-point contact electrode design, integrates a CS1253 AFE chip, supports multi-band scanning from 5kHz to 500kHz, and has an impedance measurement accuracy of ±0.5%; it has a built-in pressure sensor array to automatically detect the contact status of both feet (trigger start when contact resistance <2Ω).

[0089] The 3D volumetric mesh optical system 3 is used to acquire volumetric mesh point cloud data of user-specified parts and feed it back to the industrial control computer 5. It can use dual Intel RealSense D455 depth cameras with an 86°×57° field of view to generate a volumetric mesh model containing 50,000+ vertices (accuracy 0.5mm@1m). The specific coordinate points of the user's body volumetric mesh can be automatically tracked and monitored by the system or calibrated by the administrator, thereby realizing the construction and volume assessment of the user's body volumetric mesh model.

[0090] The millimeter-wave radar respiratory measurement unit 4 is used to scan and detect the periodic displacement change signal generated by the user's chest cavity during breathing and feed it back to the industrial control computer 5; it can be equipped with a 24GHz FMCW radar module to detect chest cavity displacement based on phase change (accuracy ±1mm) and complete 10 cycles of respiratory waveform acquisition within 60 seconds (error <6ms).

[0091] Industrial computer 5 is used to evaluate and calculate the user's BIA body composition data, respiratory rate variability index, and body grid volume based on the bioelectrical impedance data, volume grid point cloud data, and displacement periodic change signals generated by the thoracic cavity, and upload them to the edge server through the industrial gateway; the edge computing node can be based on the NVIDIA Jetson Orin Nano core, with 67 TOPS computing power, and deploy the MFN nutrition risk detection model (random forest architecture) to support real-time multimodal data fusion;

[0092] The industrial gateway is used to provide data communication services between the AI ​​nutrition screening and testing instrument and the edge server; the communication between the industrial computer, the industrial gateway and the edge server can be determined according to the gateway type selected by the user, and is not limited here;

[0093] Power supply, used for power supply; it can be solar energy, lithium battery or other power supply methods;

[0094] The foot BIA electrode module 2, the 3D volumetric mesh optical system 3, the millimeter-wave radar respiratory detection unit 4, the industrial gateway, and the power supply are all electrically connected to the industrial control computer 5.

[0095] The industrial control computer 5 communicates with the edge server through the industrial gateway.

[0096] An independent computing chip is configured in the industrial control computer, which can perform a series of preprocessing steps to filter, reduce noise, and clean the sampled data or signals. Then, based on the various types of computing models preset in the chip, the BIA body composition data, respiratory rate variability index and volume grid volume of the user are evaluated and calculated (see the calculation model of each parameter below for details).

[0097] Preferably, the BIA body composition data includes BIA muscle mass fraction, the evaluation formula of which is as follows:

[0098] ,

[0099] in,

[0100] SM represents muscle mass fraction;

[0101] H: Height (unit: centimeters), reflecting the longitudinal dimension of the human body;

[0102] R: Foot-hand bioelectrical impedance (unit: ohm), characterizing the electrical conductivity of body fluids;

[0103] S: Gender coefficient (male = 1, female = 0), used for gender difference correction;

[0104] A: Age (unit: years), considering the age-related degenerative effects of muscle.

[0105] Preferably, the calculation formula for evaluating the respiratory rate variability index is as follows:

[0106] ,

[0107] in,

[0108] SDNN represents the respiratory rate variability index;

[0109] RR i : Duration of the i-th respiratory cycle (ms);

[0110] : The average respiratory interval over N cycles;

[0111] N: Number of sampling periods (fixed to 10);

[0112] ∑: Sum of the squares of the time deviations over consecutive respiratory cycles;

[0113] √: Standard deviation calculation, quantifying the degree of fluctuation in breathing rhythm.

[0114] Preferably, the formula for evaluating the volumetric mesh volume is as follows:

[0115] ,

[0116] in,

[0117] V represents the volume of the user's body grid;

[0118] v0, v1, v2, v3: The coordinates of the tetrahedral vertices of the user's body, derived from 3D camera point cloud data;

[0119] M: The set of all tetrahedra contained in the volume mesh model;

[0120] ||: Take the absolute value to ensure that the volume is non-negative.

[0121] An industrial control computer can be configured with a controller or processing terminal, which can sample and calculate data and send it to the industrial gateway, which then reports it to the edge server.

[0122] During sampling, the industrial control computer can perform multimodal data synchronous acquisition, and use the IEEE 1588v2 protocol to achieve hardware clock synchronization (error <100ns), and align the timestamps through the ROS2 TimeSync module.

[0123] like Figure 3 The diagram shows the application system of the AI ​​nutrition screening analyzer. The AI ​​nutrition screening analyzer communicates and controls data with the edge server via an industrial gateway. The edge server can communicate with the user's app or the administrator's PC management terminal, sending the final generated nutrition test report to the user and administrator.

[0124] An MFN nutrition risk detection model is deployed on the edge server. The MFN model processing flow is as follows:

[0125] Input layer: BIA body composition data (muscle / fat ratio), respiratory rate variability, and body mesh morphological parameters (volume / curvature).

[0126] Fusion layer: Joint Representation integrates multimodal features, and the weighting coefficients are dynamically adjusted through an attention mechanism;

[0127] Output layer: Generates NRS2002 standardized nutrition risk scores and intervention recommendations.

[0128] It can achieve intelligent nutritional testing and analysis and output users' nutritional risk scores and intervention suggestions. It can realize multimodal nutritional index analysis and fusion prediction, and realize multimodal nutritional screening and testing. It has a wide range of applications and is suitable for testing different user groups. It uses AI technology to overcome the shortcomings of manual and single-computer calculation and analysis, and integrates AI technology into nutritional screening and testing technology to achieve intelligent testing results.

[0129] Based on the working principle of the aforementioned detector, and in addition, a detection method for an AI nutrition screening detector is provided, which includes the following steps:

[0130] The system initializes, establishes data communication between the AI ​​nutrition screening and testing instrument and the edge server, and constructs the user's nutrition testing report;

[0131] The edge server issues a sampling command and forwards it to the industrial control computer 5 of the Ai nutrition screening and testing instrument through the industrial gateway;

[0132] The industrial control computer 5 activates the millimeter-wave radar respiratory detection unit 4 and determines whether it has received the user's respiratory rate variability data:

[0133] If no signal is received, it indicates that there is no user or the user's posture is incorrect, and a detection alarm signal is sent to the edge server through the industrial gateway.

[0134] If received, the foot BIA electrode module 2, 3D volumetric mesh optical system 3 and millimeter-wave radar respiratory basket detection unit 4 will start sampling according to the preset sampling frequency;

[0135] The industrial control computer 5 acquires the sampling data, and based on the bioelectrical impedance data, volume mesh point cloud data and displacement periodic change signal generated by the thoracic cavity obtained from the sampling, it evaluates and calculates the user's BIA body composition data, respiratory rate variability index and volume mesh volume, and uploads them to the edge server through the industrial gateway.

[0136] The edge server receives and records the user's BIA body composition data, respiratory rate variability index, and body grid volume into the nutrition test report;

[0137] The edge server simultaneously inputs the user's BIA muscle mass score, respiratory rate variability index, and body grid volume into the preset MFN nutritional risk detection model. The MFN nutritional risk detection model identifies and outputs the corresponding nutritional risk score and intervention suggestions, and records the nutritional risk score and intervention suggestions in the nutritional detection report.

[0138] The user's nutrition test report is synchronously pushed to the APP / PC management terminal through the edge server.

[0139] The timeline of AI nutrition screening is shown in Table 1 below:

[0140] Step number Testing phase Implementing entity Key Operations Data metrics Output 1 System Initialization AI detectors / edge servers Establish a communication link and create a test report template. - Initialize the detection environment 2 Sampling command transmission Edge Server / Industrial Gateway Command routing and protocol conversion - Activate the industrial control computer 3 User Status Detection Industrial PC 5 Millimeter-wave radar attitude detection respiratory rate variability User status verification / alarm signal 4 Multimodal data acquisition sensor array Simultaneous acquisition of bioimpedance / volumetric / respiratory data BIA data, 3D point cloud, chest cavity displacement Original dataset 5 Feature calculation Industrial PC 5 Body composition analysis / volume calculation / respiratory assessment Muscle mass fraction, volume grid volume Structured feature data 6 Risk Prediction MFN model Multi-feature fusion analysis Nutritional risk score Intervention Recommendation Report 7 Model training Machine learning systems Historical data feature engineering Clinical assessment association features Validated prediction model 8 Data push Kafka middleware Cross-device data synchronization - Accessible electronic reports to APP / PC

[0141] Table 1

[0142] In step 3, the present invention needs to detect whether the user has reached the detection position, that is, whether the user is standing in the correct position. The user's chest cavity is detected by millimeter-wave radar attitude detection scanning, and the chest cavity displacement signal is detected based on the phase change. If a valid signal is received, it indicates that the user's posture is correct (the specific correct posture can be corrected and identified by the server in advance). If no signal is received or an invalid signal is received, an alarm prompt is issued.

[0143] Preferably, the method for generating the MFN nutritional risk detection model includes:

[0144] Collect and preprocess historical nutritional consultation big data from several users. The historical nutritional consultation big data includes users' BIA body composition data, respiratory rate variability index and body grid volume, as well as nutritional risk scores determined by quantitative assessment forms and intervention recommendations issued by nutritionists during clinical consultations.

[0145] Feature engineering was performed on the historical nutritional visit big data to extract corresponding BIA body composition features, respiratory rate variability features and volume grid volume features, and feature fusion was performed to generate historical nutritional visit features for each user.

[0146] The historical nutritional consultation characteristics of each user, as well as the corresponding nutritional risk scores and intervention suggestions, are statistically analyzed to generate a feature set consisting of several {historical nutritional consultation characteristics, nutritional risk scores, and intervention suggestions}, which are then divided into a training set and a validation set according to a preset ratio.

[0147] The training set is input into the preset MFN model to perform feature classification learning and generate an initial MFN nutrition risk detection model.

[0148] The validation set is input into the MFN nutrition risk detection model to verify its predictive and identification performance.

[0149] If the verification is successful, the MFN nutrition risk detection model is deployed and applied to the edge server;

[0150] Conversely, repeat the above steps to reconstruct the MFN nutritional risk detection model.

[0151] The construction process of the MFN nutrition risk detection model is shown in the timeline table in Table 2 below:

[0152] Step number Steps Specific content Technical points 1 Data Acquisition and Preprocessing We collected historical data from over 10,000 users, including: · BIA body composition data (preferably muscle mass) · Respiratory rate variability (RRV ≥ 3.5ms as the outlier threshold) · 3D volumetric mesh data (accuracy ±50ml) · NRS-2002 nutritional risk score (0-7 point scale) · Physician intervention recommendations (dietary / enteral / parenteral nutrition). Preprocessing included data cleaning (removing outliers > 5σ) and format standardization (Z-score normalization). Establish a multimodal data warehouse and implement HIPAA compliant encryption. 2 Feature engineering construction Three main feature systems were extracted: 1. BIA muscle mass feature (muscle mass value); 2. Respiratory rate variability feature (respiratory rate variability index value); 3. Volume mesh geometric features (chest-to-abdomen volume ratio / limb volume ratio or other body volume mesh volume values). A 128-dimensional feature vector was generated through a feature fusion layer. Fusion using PCA dimensionality reduction + Attention mechanism 3 Dataset Construction Generate structured feature sets: • Training set (n=8000): 7:3 stratified sampling • Validation set (n=2000): Includes seasonal fluctuation data • Test set (n=1000): An independent external dataset. Annotation dimensions include nutritional risk level (low / medium / high) and 22 types of interventions. SMOTE oversampling processing category imbalance 4 Model training Constructing the MFN multi-branch network: • BIA feature branch: 3-layer CNN • Breathing feature branch: LSTM network • Volume feature branch: Graph convolutional network Joint training employs a Dynamic Weight Adjustment (DWA) strategy. Adam optimizer (lr=0.001) + FocalLoss 5 Model Validation Evaluation metrics include: AUC-ROC ≥ 0.85; Risk grading accuracy > 80%; Intervention recommendation matching accuracy F1-score > 0.75; Deployment standard requires inference latency < 300ms (NVIDIA Jetson edge device). SHAP value interpretability analysis 6 Iterative optimization If the target is not met, initiate the optimization process: 1. Expand the high-risk population sample (BMI > 30 or albumin < 35g / L) 2. Add dynamic features (continuous monitoring data) 3. Improve the fusion layer structure (gated attention mechanism) Transfer learning (MedicalNet pre-training)

[0153] Table 2

[0154] The MFN model architecture described above will not be repeated in this embodiment. Feature engineering can be performed using data statistical algorithms or text recognition algorithms to extract corresponding feature values ​​or keywords. Specifically, this can be determined by statistically analyzing the detection values ​​of each user.

[0155] The following are application examples of the above model:

[0156] Example 1

[0157] Patient AI Nutritional Screening Test Report 1

[0158] 1) Patient Information:

[0159] Zhang San, male, 58 years old, 172cm tall, 75kg;

[0160] 2) Test data:

[0161] BIA body composition data: body fat percentage 22% (normal), skeletal muscle mass 28kg (normal), basal metabolic rate 1580kcal;

[0162] Respiratory rate variability: mean respiratory rate 14 breaths / min, variability index 0.8 (normal range).

[0163] Body grid volume: The standard value for volume corresponding to height is 1.02 (without edema or muscle atrophy);

[0164] 3) MFN model output:

[0165] Nutritional risk score: 2 points (low risk);

[0166] Intervention recommendations: Maintain your current balanced diet and have your body composition and metabolic rate checked monthly.

[0167] Example 2

[0168] Patient AI Nutritional Screening Test Report 2

[0169] 1) Patient Information:

[0170] Li Si, female, 72 years old, 160cm tall, 48kg;

[0171] 2) Test data

[0172] BIA body composition data: body fat percentage 15% (low), skeletal muscle mass 18kg (below standard), basal metabolic rate 1120kcal;

[0173] Respiratory rate variability index: mean respiratory rate 20 breaths / min, variability index 2.5 (abnormally high, indicating respiratory compensation);

[0174] Body grid volume: The standard value for volume corresponding to height is 0.89 (indicating reduced muscle mass or malnutrition);

[0175] 3) MFN model output

[0176] Nutritional risk score: 7 (high risk);

[0177] Intervention recommendations: Immediately supplement with high-protein nutritional formulas, monitor respiratory function, and it is recommended to combine clinical examinations to rule out chronic wasting diseases. Nutritional indicators should be retested weekly.

[0178] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 4 As shown, optionally, electronic device 410 may include a first processor 2001.

[0179] Optionally, the electronic device 410 may also include a memory 2002 and a transceiver 2003.

[0180] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0181] The following is combined with Figure 4 A detailed description of each component of electronic device 410 is provided below:

[0182] The first processor 2001 is the control center of the electronic device 410. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0183] Optionally, the first processor 2001 can perform various functions of the electronic device 410 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0184] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 4 CPU0 and CPU1 are shown in the diagram.

[0185] In a specific implementation, as one example, the electronic device 410 may also include multiple processors, for example... Figure 4 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0186] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0187] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently and be connected via the interface circuit of the electronic device 410. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0188] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0189] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 4 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0190] Optionally, the transceiver 2003 can be integrated with the first processor 2001, or it can exist independently and be connected via the interface circuit of the electronic device 410. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0191] It should be noted that, Figure 4 The structure of the electronic device 410 shown does not constitute a limitation on the router. Actual knowledge structure identification devices may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0192] Furthermore, the technical effects of the electronic device 410 can be referenced from the technical effects of the detection method of the Ai nutrition screening detector described in the above method embodiments, and will not be repeated here.

[0193] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0194] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0195] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0196] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0197] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0198] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0199] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0200] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0201] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0202] The units described as separate components may or may not be physically separate. 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 units can be selected to achieve the purpose of this embodiment according to actual needs.

[0203] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0204] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0205] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An AI nutrition screening and testing instrument, the AI ​​nutrition screening and testing instrument comprising a testing platform (1) and a support (6) disposed on the testing platform (1), characterized in that, The Ai nutrition screening and testing instrument also includes: Foot BIA electrode module (2) is used to acquire bioelectrical impedance data of the user's foot and feed it back to the industrial control computer (5); 3D volumetric mesh optical system (3) is used to acquire volumetric mesh point cloud data of user-specified parts and feed it back to the industrial control computer (5); The millimeter-wave radar respiratory basket detection unit (4) is used to scan and detect the periodic displacement change signal generated by the user's chest cavity when breathing and feed it back to the industrial control computer (5). The industrial control computer (5) is used to evaluate and calculate the user's BIA body composition data, respiratory rate variability index and body grid volume based on the bioelectrical impedance data, body grid point cloud data and displacement periodic change signal generated by the thoracic cavity, and upload them to the edge server through the industrial gateway. An industrial gateway is used to provide data communication services between the AI ​​nutrition screening and testing instrument and the edge server; Power supply, used for supplying power; The foot BIA electrode module (2), 3D volumetric mesh optical system (3), millimeter-wave radar respiratory detection unit (4), industrial gateway and power supply are electrically connected to the industrial control computer (5), respectively. The industrial control computer (5) communicates with the edge server through the industrial gateway.

2. The AI ​​nutrition screening and testing instrument according to claim 1, characterized in that, The BIA body composition data includes BIA muscle mass fraction, which is calculated using the following formula: ， in, SM represents muscle mass fraction; H: Height (unit: centimeters), reflecting the longitudinal dimension of the human body; R: Foot-hand bioelectrical impedance (unit: ohm), characterizing the electrical conductivity of body fluids; S: Gender coefficient (male = 1, female = 0), used for gender difference correction; A: Age (unit: years), considering the age-related degenerative effects of muscle.

3. The AI ​​nutrition screening and testing instrument according to claim 1, characterized in that, The formula for evaluating the respiratory rate variability index is as follows: ， in, SDNN represents the respiratory rate variability index; RR i : Duration of the i-th respiratory cycle (ms); : The average respiratory interval over N cycles; N: Number of sampling periods (fixed to 10); ∑: Sum of the squares of the time deviations over consecutive respiratory cycles; √: Standard deviation calculation, quantifying the degree of fluctuation in breathing rhythm.

4. The AI ​​nutrition screening and testing instrument according to claim 1, characterized in that, The formula for evaluating the volumetric mesh volume is as follows: ， in, V represents the volume of the user's body grid; v0, v1, v2, v3: The coordinates of the tetrahedral vertices of the user's body, derived from 3D camera point cloud data; M: The set of all tetrahedra contained in the volume mesh model; ||: Take the absolute value to ensure that the volume is non-negative.

5. A detection method for the Ai nutrition screening and detection instrument according to any one of claims 1-4, characterized in that, The detection method includes the following steps: The system initializes, establishes data communication between the AI ​​nutrition screening and testing instrument and the edge server, and constructs the user's nutrition testing report; The edge server issues a sampling command and forwards it to the industrial control computer (5) of the Ai nutrition screening and testing instrument through the industrial gateway. The industrial control computer (5) starts activating the millimeter-wave radar respiratory basket detection unit (4) and determines whether it has received the user's respiratory rate variability index data: If no signal is received, it indicates that there is no user or the user's posture is incorrect, and a detection alarm signal is sent to the edge server through the industrial gateway. If received, the foot BIA electrode module (2), 3D volumetric optical system (3) and millimeter-wave radar respiratory basket measurement unit (4) will start sampling according to the preset sampling frequency; The industrial control computer (5) acquires the sampling data, and evaluates and calculates the user's BIA body composition data, respiratory rate variability index and body grid volume based on the bioelectrical impedance data, volume grid point cloud data and displacement periodic change signal generated by the thoracic cavity obtained by sampling, and uploads them to the edge server through the industrial gateway. The edge server receives and records the user's BIA body composition data, respiratory rate variability index, and body grid volume into the nutrition test report; The edge server simultaneously inputs the user's BIA muscle mass score, respiratory rate variability index, and body grid volume into the preset MFN nutritional risk detection model. The MFN nutritional risk detection model identifies and outputs the corresponding nutritional risk score and intervention suggestions, and records the nutritional risk score and intervention suggestions in the nutritional detection report.

6. The detection method according to claim 5, characterized in that, The detection method further includes the following steps: The user's nutrition test report is synchronously pushed to the APP / PC management terminal through the edge server.

7. The Ai nutrition screening instrument and detection method according to claim 5, characterized in that, The method for generating the MFN nutritional risk detection model includes: Collect and preprocess historical nutritional consultation big data from several users. The historical nutritional consultation big data includes users' BIA body composition data, respiratory rate variability index and body grid volume, as well as nutritional risk scores determined by quantitative assessment forms and intervention recommendations issued by nutritionists during clinical consultations. Feature engineering was performed on the historical nutritional visit big data to extract corresponding BIA body composition features, respiratory rate variability features and volume grid volume features, and feature fusion was performed to generate historical nutritional visit features for each user. The historical nutritional consultation characteristics of each user, as well as the corresponding nutritional risk scores and intervention suggestions, are statistically analyzed to generate a feature set consisting of several {historical nutritional consultation characteristics, nutritional risk scores, and intervention suggestions}, which are then divided into a training set and a validation set according to a preset ratio. The training set is input into the preset MFN model to perform feature classification learning and generate an initial MFN nutrition risk detection model. The validation set is input into the MFN nutrition risk detection model to verify its predictive and identification performance. If the verification is successful, the MFN nutrition risk detection model is deployed and applied to the edge server; Conversely, repeat the above steps to reconstruct the MFN nutritional risk detection model.

8. An electronic device, characterized in that, The electronic device includes: processor; A memory storing computer-readable instructions that, when executed by the processor, implement the method as described in any one of claims 5 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 5 to 7.

Images