Intelligent obesity assessment and personalized weight loss management system fusing multi-modal data

Abstract

This invention discloses an intelligent obesity assessment and personalized weight loss management system that integrates multimodal data, belonging to the field of intelligent health management technology. It includes: a hierarchical data acquisition module configured to collect general public data and specialized medical data hierarchically through wearable device interfaces, manual input interfaces, and third-party data access interfaces. The general data includes height, weight, dietary records, and exercise records, while the specialized medical data includes complication diagnosis results, physical examination report data, and clinical treatment records. A dual-model data analysis module is also included, incorporating a clinical in-depth analysis model and a public health model. This invention solves the problems of existing systems having limited adaptability to specific scenarios, poor data integration compatibility, and insufficient security and compliance. It not only meets the general weight loss needs of the general public but also provides clinically adapted diagnostic and treatment support for specialists, achieving precise and compliant management of obesity across all scenarios.

Description

Technical Field

[0001] This invention belongs to the field of intelligent health management technology, specifically involving an intelligent obesity assessment and personalized weight loss management system that integrates multimodal data. Background Technology

[0002] As the global obesity problem becomes increasingly severe, obesity has become a core contributing factor to various chronic diseases such as fatty liver, diabetes, and cardiovascular and cerebrovascular diseases, affecting not only the quality of life of ordinary people but also placing a heavy burden on the healthcare system. Existing obesity management systems have significant shortcomings: most systems only focus on the general weight loss needs of the population, providing only basic dietary and exercise advice, and cannot meet the specialist treatment needs of patients with complications; some medical systems are only geared towards clinical scenarios, are complex to operate, and lack general applicability, making it difficult to cover the daily management needs of the general public.

[0003] Meanwhile, existing technologies also suffer from problems such as poor data integration compatibility and insufficient security compliance: general data and specialized medical data have different formats, making it difficult to achieve effective fusion analysis; there is a risk of privacy leakage during the transmission and storage of medical data, which does not comply with medical data security management standards; and the model training does not fully incorporate clinical guidelines, resulting in insufficient medical adaptability of the output results.

[0004] Therefore, there is an urgent need for an intelligent management system that can take into account both public needs and professional standards, and achieve accurate data collection, secure integration, and dual-scenario adaptive analysis.

[0005] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent obesity assessment and personalized weight loss management system that integrates multimodal data. This system can solve problems such as the inability of existing systems to simultaneously meet the general weight loss needs of the general public and the requirements of specialized medical treatment, limited scenario adaptability, poor compatibility between general data and specialized medical data, lack of standardized processing mechanisms, insufficient medical data security and compliance, and low clinical adaptability of model analysis results.

[0007] To achieve the above objectives, a specific embodiment of the present invention provides the following technical solution: An intelligent obesity assessment and personalized weight loss management system integrating multimodal data includes: The layered data acquisition module is configured to collect general public data and specialized medical data in layers through wearable device interface, manual input interface and third-party data access interface. The general data includes height, weight, diet records, exercise records and sleep habits, while the specialized medical data includes complication diagnosis results, physical examination report data and clinical diagnosis and treatment records. The dual-model data analysis module incorporates a clinical in-depth analysis model and a public health model. The clinical in-depth analysis model is based on a multimodal data fusion engine, integrating convolutional neural networks (CNN) for structured feature extraction from medical images and recurrent neural networks (RNN) for dynamic analysis of time-series physiological data. It is also deeply embedded in the rule base of guidelines for the diagnosis and treatment of obesity and related metabolic diseases, enabling the classification of obesity etiology, quantitative assessment of the severity of complications, and evidence-based recommendations for intervention programs. The public health model analyzes general data based on decision tree and random forest algorithms, outputting weight trend prediction results and popularized diet and exercise programs. The dual-end adaptive interaction module includes a public-end interaction interface and a specialist-end interaction interface. The public-end interaction interface is used for general data entry, weight loss progress display, and viewing of popular plans. The specialist-end interaction interface is used for specialist medical data entry, clinical data query, and follow-up plan formulation. The medical-grade security module is configured to employ transmission encryption, storage encryption, multi-factor authentication, and fine-grained authorization management mechanisms to ensure both general data privacy and the security and compliance of specialized medical data. The hybrid cloud integration module adopts a hybrid cloud architecture, with general data deployed on the public cloud and specialized medical data deployed on a private cloud or a hospital-specific cloud node. It achieves integration and docking with external systems through an API gateway.

[0008] In one or more embodiments of the present invention, in the layered data acquisition module, the wearable device interface is adapted to Bluetooth BLE 5.0 and above communication protocols, and is compatible with smart bracelets, body fat scales, and medical-grade dynamic blood glucose monitors, for real-time transmission of step count, body fat percentage sub-indicators, resting metabolic rate, and blood glucose fluctuation curve data.

[0009] In one or more embodiments of the present invention, the third-party data access interface of the hierarchical data acquisition module includes a general interface and a medical interface. The general interface connects to health apps and fitness platforms, and the medical interface uses the HL7FHIR protocol to connect to hospital HIS / EMR / LIS systems. Furthermore, a "general data dictionary + specialty data dictionary" is established to achieve cross-source data standardization.

[0010] In one or more embodiments of the present invention, the clinical deep analysis model has a built-in guideline rule base that supports precise stratification of complications based on the joint determination of multiple indicators. It embeds specialty rules in the diagnosis and treatment guidelines for the precise identification of obesity with diabetes, obesity with fatty liver, obesity with abnormal blood lipids and blood pressure, and obesity with obstructive sleep apnea syndrome (OSA).

[0011] In one or more embodiments of the present invention, the medical-grade security module uses SSL / TLS protocol to encrypt general data and HTTPS+SM4 encryption protocol to encrypt specialized medical data for transmission encryption; storage encryption uses an "encrypted partition + access permission binding" mechanism for specialized medical data, which can only be decrypted and viewed by authorized doctors.

[0012] In one or more embodiments of the present invention, the fine-grained authorization management mechanism of the medical-grade security module includes a four-level permission system: users can only view their own general data, nutritionists can view users' diet and exercise data, doctors can view all user data, and administrators are only responsible for system configuration and have no data access permissions.

[0013] In one or more embodiments of the present invention, the API gateway of the hybrid cloud integration module includes a general API gateway and a dedicated medical API gateway. The dedicated medical API gateway separately manages interface access with the hospital system and sets traffic limits and security verification.

[0014] In one or more embodiments of the present invention, the dual-end adaptation interaction module adopts a responsive design and adapts to computers, mobile phones, and tablet terminals based on media queries. The specialist end interaction interface supports quick selection of medical indicators and voice input of medical record summaries.

[0015] In one or more embodiments of the present invention, the dual-model data analysis module is further configured with a stream computing framework for real-time processing of dynamic data collected by wearable devices, and for pushing exercise target reminders to the public and abnormal patient indicator warnings to specialists.

[0016] In one or more embodiments of the present invention, the hybrid cloud integration module can interface with the nutrition clinic system, the bariatric surgery center system, and the rehabilitation institution system to achieve bidirectional synchronization of specialist treatment plans and home management data.

[0017] Compared with existing technologies, the intelligent obesity assessment and personalized weight loss management system that integrates multimodal data of the present invention has the following beneficial effects: (1) It has achieved full coverage of both "daily weight loss management for the general public" and "precision diagnosis and treatment intervention for specialist patients", solving the problem of single scenario adaptation of the existing system; (2) Through hierarchical data collection and dual-model parallel analysis architecture, general data and specialized medical data are effectively integrated, which significantly improves the accuracy of obesity assessment and the personalization and clinical adaptability of intervention programs. (3) Adopting a medical-grade security protection mechanism and hybrid cloud architecture, it meets the requirements of laws and regulations such as the Personal Information Protection Law and the Information Security Management Standard for Medical and Health Institutions, ensuring data security and privacy compliance; (4) It supports integration with multiple systems such as hospital system, nutrition department, and bariatric surgery, and opens up the whole chain of "clinical diagnosis and treatment - home management - rehabilitation follow-up", realizing the continuity and closed loop of obesity management. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A schematic diagram of the core business process of a specialized, adaptive intelligent obesity management system that integrates multimodal data. Detailed Implementation

[0020] To enable those skilled in the art to better understand the technical solutions in this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this disclosure.

[0021] like Figure 1 As shown, an embodiment of the intelligent obesity assessment and personalized weight loss management system integrating multimodal data according to the present invention includes a hierarchical data acquisition module, a dual-model data analysis module, a dual-end adaptation and interaction module, a medical-grade security module, and a hybrid cloud integration module. Specific implementation details are as follows: The hierarchical data acquisition module is implemented as follows: Wearable device interface development can utilize Python's pybluez library to implement Bluetooth communication protocol adaptation, ensuring compatibility with smart bracelets, body fat scales, and continuous glucose monitors. Data preprocessing is divided into two modes: for general data, such as step counts, a mean filtering algorithm is used to remove noise; for specialized data, such as blood glucose curves, a linear calibration algorithm is used for precise correction, preserving key clinical feature points, such as the 2-hour postprandial blood glucose peak.

[0022] The manual data entry interface for the general public is developed using a modular approach, integrating a food image recognition model to automatically estimate calories; the specialist interface has preset clinical fields that support matching with ICD-10 codes and has built-in verification rules that automatically mark abnormal data as "requiring doctor's review".

[0023] In the third-party data access interface, the medical interface connects with the hospital's HIS / EMR / LIS system via the HL7 FHIR standard protocol, and data anonymization is performed before transmission. Semantic alignment and standardization of data from different sources are achieved through the establishment of a "general data dictionary" and a "specialty data dictionary" (following standards such as the "Obesity Diagnosis and Treatment Guidelines").

[0024] The dual-model data analysis module is implemented as follows: Popular health model: Built on the Scikit-learn framework, it uses algorithms such as decision trees and random forests, and is trained using common data of ordinary obese users. The model outputs weight trend predictions and popular diet and exercise plans.

[0025] A deep clinical analysis model is built using the TensorFlow / PyTorch framework. The CNN submodule extracts textual features from physical examination reports and spatial features from images such as liver ultrasound; the RNN submodule analyzes dynamic patterns in time-series physiological data such as blood glucose and blood pressure. The model is deeply embedded in a structured clinical guideline rule knowledge base. It is trained using ethically reviewed and anonymized data from patients with obesity-related complications, supplemented with publicly available datasets to cover rare types. Model outputs include obesity classification, complication risk levels, and drug / surgical evaluation recommendations. It can automatically generate a structured "Intelligent Obesity Assessment and Intervention Recommendation Report," which can be directly imported into hospital electronic medical record systems.

[0026] Real-time stream processing: Integrates stream computing frameworks such as Apache Flink to process wearable device data in real time. The user side can update progress periodically; the specialist side can set clinical thresholds, such as fasting blood glucose ≥7.0 mmol / L, and send real-time alerts to doctors when triggered.

[0027] The dual-platform adaptation and interaction module is implemented as follows: The user interface adopts a fresh and lively design style, visually displaying progress through line charts, pie charts, and other visual aids, and setting achievement badges for incentives.

[0028] The specialist interface adopts a professional and concise medical style, highlights key indicators with colors, and provides professional charts such as blood glucose curves and body fat-complication correlation heatmaps, and supports one-click export to medical records.

[0029] Employing a responsive design based on CSS media queries, it adapts to computers, tablets, and mobile phones. The specialist app additionally provides a quick selection list of medical indicators and voice input for medical record summaries, improving doctors' work efficiency.

[0030] The medical-grade security module is implemented as follows: Encryption of transmission: General data transmission uses the SSL / TLS 1.3 protocol; specialized medical data transmission uses HTTPS with the national cryptographic SM4 algorithm for encryption.

[0031] Storage encryption: Specialty medical data is encrypted using the AES-256 algorithm and a "encrypted partition + access permission binding" mechanism is implemented to ensure that only authorized doctors can decrypt and access the data.

[0032] Access control: Implement a four-level fine-grained permission system (user, nutritionist, doctor, administrator). For high-privilege users such as doctors, enforce multi-factor authentication, such as employee ID + password + UKey.

[0033] Security audit: Record all access and operation logs to specialized data, and monitor abnormal behaviors in real time, such as batch exporting outside of working hours, and trigger security alerts in a timely manner.

[0034] Hybrid cloud integration module, the specific implementation of which is as follows: Architecture Deployment: A hybrid cloud architecture combining public cloud and hospital private cloud is adopted. General data and services are deployed on the public cloud to ensure elasticity and cost; specialized medical data and clinical models are deployed on the hospital's private cloud or dedicated cloud nodes that have passed the Level 3 Information Security Protection Certification to ensure data sovereignty and the highest level of security.

[0035] API Gateway: Built using Spring Cloud Gateway and other technologies. A dedicated gateway for medical APIs is established to connect to the hospital's internal systems, implementing strict traffic control and security checks, such as a traffic control limit of 50 requests per second.

[0036] System Integration: Through standard interfaces, it connects with hospital HIS, nutrition outpatient systems, and bariatric surgery center systems to achieve two-way synchronization of treatment plans, home data, rehabilitation programs, and billing information, building an interconnected management ecosystem.

[0037] This invention constructs an intelligent obesity management system that caters to both general public and specialist needs through five core modules: hierarchical data acquisition, dual-model parallel analysis, dual-platform adaptive interaction, medical-grade security protection, and hybrid cloud integration. The hierarchical data acquisition module achieves accurate collection and standardized integration of general and specialized medical data; the dual-model data analysis module adapts to both daily management and clinical treatment scenarios, outputting targeted results; the dual-platform adaptive interaction module optimizes the user experience for different users; the medical-grade security module ensures data privacy and compliance; and the hybrid cloud integration module enables multi-system integration and flexible deployment, ultimately achieving full-scenario coverage of "daily weight loss management for the general public + precise treatment intervention for specialist patients."

[0038] It will be apparent to those skilled in the art that this disclosure is not limited to the details of the exemplary embodiments described above, and that this disclosure can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of this disclosure is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within this disclosure. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0039] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. An intelligent obesity assessment and personalized weight loss management system integrating multimodal data, characterized in that, include: The layered data acquisition module is configured to collect general public data and specialized medical data in layers through wearable device interface, manual input interface and third-party data access interface. The general data includes height, weight, diet records, exercise records and sleep habits, while the specialized medical data includes complication diagnosis results, physical examination report data and clinical diagnosis and treatment records. The dual-model data analysis module incorporates a clinical in-depth analysis model and a public health model. The clinical in-depth analysis model is based on a multimodal data fusion engine, integrating convolutional neural networks (CNNs) for structured feature extraction from medical images and recurrent neural networks (RNNs) for dynamic analysis of time-series physiological data. It is also deeply embedded in the rule base of guidelines for the diagnosis and treatment of obesity and related metabolic diseases, enabling the classification of obesity etiology, quantitative assessment of the severity of complications, and evidence-based recommendations for intervention programs. The public health model analyzes general data based on decision tree algorithms and random forest algorithms, outputting weight trend prediction results and popularized diet and exercise programs. The dual-end adaptive interaction module includes a public-end interaction interface and a specialist-end interaction interface. The public-end interaction interface is used for general data entry, weight loss progress display, and viewing of popular plans. The specialist-end interaction interface is used for specialist medical data entry, clinical data query, and follow-up plan formulation. The medical-grade security module is configured to employ transmission encryption, storage encryption, multi-factor authentication, and fine-grained authorization management mechanisms to ensure both general data privacy and the security and compliance of specialized medical data. The hybrid cloud integration module adopts a hybrid cloud architecture, with general data deployed on the public cloud and specialized medical data deployed on a private cloud or a hospital-specific cloud node. It achieves integration and docking with external systems through an API gateway.

2. The system according to claim 1, characterized in that, In the layered data acquisition module, the wearable device interface is compatible with Bluetooth BLE 5.0 and above communication protocols, and is compatible with smart bracelets, body fat scales, and medical-grade dynamic blood glucose monitors, for real-time transmission of step count, body fat percentage sub-indicators, resting metabolic rate, and blood glucose fluctuation curve data.

3. The system according to claim 1, characterized in that, The third-party data access interface of the hierarchical data acquisition module includes a general interface and a medical interface. The general interface connects to health apps and fitness platforms, while the medical interface uses the HL7 FHIR protocol to connect to hospital HIS / EMR / LIS systems and establishes a "general data dictionary + specialty data dictionary" to achieve cross-source data standardization.

4. The system according to claim 1, characterized in that, The clinical deep analysis model has a built-in guideline rule base that supports precise stratification of complications based on the joint determination of multiple indicators. It embeds specialty rules in the diagnosis and treatment guidelines for the accurate identification of obesity with diabetes, obesity with abnormal blood lipids and blood pressure, obesity with fatty liver, and obesity with obstructive sleep apnea syndrome (OSA).

5. The system according to claim 1, characterized in that, In the medical-grade security module, transmission encryption uses the SSL / TLS protocol to encrypt general data, and the HTTPS+SM4 encryption protocol to encrypt specialized medical data; storage encryption uses an "encrypted partition + access permission binding" mechanism for specialized medical data, which can only be decrypted and viewed by authorized doctors.

6. The system according to claim 1, characterized in that, The fine-grained authorization management mechanism of the medical-grade security module includes a four-level permission system: users can only view their own general data, nutritionists can view users' diet and exercise data, doctors can view all user data, and administrators are only responsible for system configuration and have no data access permissions.

7. The system according to claim 1, characterized in that, The API gateway of the hybrid cloud integration module includes a general API gateway and a dedicated medical API gateway. The dedicated medical API gateway separately manages interface access with the hospital system and sets traffic limits and security verification.

8. The system according to claim 1, characterized in that, The dual-end adaptive interaction module adopts a responsive design and is adapted to computers, mobile phones, and tablets based on media queries. The specialist terminal interaction interface supports quick selection of medical indicators and voice input of medical record summaries.

9. The system according to claim 1, characterized in that, The dual-model data analysis module is also equipped with a stream computing framework, which is used to process dynamic data collected by wearable devices in real time, and push exercise target reminders to the public and abnormal patient indicators to specialists.

10. The system according to claim 1, characterized in that, The hybrid cloud integration module can interface with the nutrition clinic system, the bariatric surgery center system, and the rehabilitation institution system to achieve two-way synchronization of specialist treatment plans and home management data.

Images