Safe and interoperable multi-part microcirculation data management method and system

Abstract

The invention discloses a safe and interoperable multi-part microcirculation data management method and system, and the method comprises the steps: designing a data model of a standard data structure based on an FHIR standard; preprocessing the microcirculation data to obtain preprocessed data; based on a hybrid encryption technology, the transmission and storage process of the preprocessed data is encrypted, and ciphertext data is obtained; performing data backup and recovery on the ciphertext data based on a multi-level backup and recovery technology to obtain decrypted data; and evaluating and predicting the health state of the user based on the multi-modal prediction model and the decrypted data. According to the invention, standardized storage, sharing and efficient analysis of medical data are realized. By integrating nail fold, sole and retina microcirculation images, dynamic blood flow monitoring data and medical text records, the system supports unified management and deep analysis of multi-modal data, and the accuracy of health management and disease prediction is improved.

Description

Technical Field

[0001] This application belongs to the technical field of data management, and particularly relates to a secure and interoperable multi-site microcirculation data management method and system. Background Art

[0002] In recent years, the importance of multi-site microcirculation analysis has become increasingly prominent. For example, as mentioned in the paper "Research Progress of Nailfold Microcirculation Detection in Connective Tissue Diseases", nailfold microcirculation detection is a non-invasive and intuitive imaging technique that can quantify microvascular lesions and has been included in the 2013 SSc classification criteria. Its functional status is closely related to connective tissue diseases and others. For example, as mentioned in the paper "Retinal Microcirculation Detection Technology and Its Clinical Applications", retinal microcirculation analysis can be used to detect fundus vascular diseases such as retinal lesions and choroid lesions through an ophthalmoscope, and has important clinical significance for the early identification and precise treatment of hypertensive eye target organ damage. For example, as mentioned in the "Guidelines for Multidisciplinary Approaches to the Prevention and Management of Diabetic Foot Disease (2020 Edition)", plantar microcirculation analysis also has potential value for evaluating diseases such as lower extremity vascular health and diabetic foot. At the same time, the integrated analysis of nailfold, plantar, and retinal microcirculation is also of great significance for the early detection and monitoring of diseases such as diabetes. Currently, there is an urgent need for a data management method to centrally store and manage multi-site microcirculation data and perform integrated analysis, providing stronger data support for the early diagnosis and precise management of diseases. However, how to effectively manage and integrate analyze these massive data to improve the scientificity and accuracy of clinical decision-making has always been a research hotspot in the field of medical informatics.

[0003] The effective utilization of microcirculation data depends on the collaboration among multiple medical institutions. However, traditional data management methods often lack unified data standards and efficient information sharing mechanisms, resulting in the inability of data to flow smoothly between institutions. This "data island" phenomenon makes it difficult for doctors to comprehensively understand the microcirculation status of patients, thus affecting the accuracy of clinical decision-making. Moreover, different medical institutions use different data storage systems, leading to obstacles in the integration and unified analysis of microcirculation data.

[0004] Traditional data management methods usually adopt a centralized storage architecture. The characteristic of centralized data storage makes it a major target of cyberattacks. Once the storage server is attacked, the microcirculation data of patients and other sensitive information will face the risk of leakage or being maliciously tampered with. This not only poses a threat to patients' privacy but may also lead to serious legal and social problems. For example, in the paper "Research on Ethical Considerations and Strategies in the Application of Health Information Technology", it is emphasized that patients' right to privacy is one of the core contents of medical ethics, and any unauthorized data access and use may cause irreparable losses to patients. Moreover, traditional data management methods lack perfect encryption and authorization mechanisms during data transmission and storage, resulting in patients' data being easily illegally accessed or used by unauthorized parties, seriously threatening patients' privacy rights.

[0005] Microcirculation data has characteristics such as multi-dimension, high frequency, and heterogeneity. The limitations of traditional data management methods in data processing and storage capabilities lead to a large amount of data not being efficiently utilized. For example, the paper "Medical Healthcare Blockchain IoMT System for Privacy Protection and Fraud Support Based on Federated Learning" discusses the defects of traditional data management methods in the field of healthcare, especially the problems in data security, information sharing, privacy protection, data management efficiency, etc. In addition, traditional data management methods lack intelligent data analysis and management functions and are difficult to support real-time monitoring, analysis, and mining of large-scale microcirculation data.

[0006] In summary, traditional data management methods have many drawbacks such as insufficient data security, difficult information sharing, insufficient privacy protection, low data management efficiency, and insufficient data integration and analysis. Against this background, an emerging hierarchical data management method based on lightweight blockchain and federated learning is proposed to solve the above problems to improve the accuracy, comprehensiveness, and efficiency of multi-site microcirculation data management, providing stronger data support for early disease diagnosis and precise management. Summary of the Invention

[0007] This application provides a secure and interoperable multi-site microcirculation data management method and system to solve the technical problems of traditional data management methods, such as insufficient data security, difficult information sharing, insufficient privacy protection, low data management efficiency, and insufficient data integration and analysis.

[0008] To solve the above technical problems, a technical solution adopted by this application is: A secure and interoperable multi-site microcirculation data management method, including:

[0009] Based on the FHIR standard, design a data model with a standard data structure to obtain microcirculation data;

[0010] Preprocess the microcirculation data to obtain preprocessed data;

[0011] Based on the hybrid encryption technology, encrypt the transmission and storage processes of the preprocessed data to obtain ciphertext data;

[0012] Based on the multi-level backup and recovery technology, perform data backup and recovery on the ciphertext data to obtain decrypted data;

[0013] Based on the multi-modal prediction model and the decrypted data, evaluate and predict the user's health status.

[0014] Furthermore, the method for designing the data structure includes:

[0015] Select and customize FHIR resources based on the core functions of the data structure;

[0016] Based on the FHIR resources, define the relationships between FHIR resources to obtain standard resources;

[0017] Based on the HAPI FHIR framework and the standard data, perform data storage and interface design to obtain microcirculation data.

[0018] Furthermore, the method for obtaining ciphertext data includes:

[0019] Generate a random symmetric first encryption key;

[0020] Based on the asymmetric encryption key, encrypt the first encryption key to obtain a second encryption key;

[0021] Based on the second encryption key, encrypt the preprocessed data to obtain ciphertext data;

[0022] Store the second encryption key and the ciphertext data in the blockchain, generate a hash fingerprint of the ciphertext digital data and store it in the blockchain to ensure the immutability of the data.

[0023] Furthermore, the method for performing data backup and recovery on the ciphertext data includes:

[0024] Evaluate the importance of the ciphertext data to determine the backup strategy combination; among them, the backup strategy combination includes full backup, incremental backup, and differential backup;

[0025] Based on the backup strategy combination, perform a backup operation on the ciphertext data to obtain backup data;

[0026] Regularly perform a recovery test on the backup data, simulate the data recovery process, and verify the availability of the backup for the backup data and the effectiveness of the recovery process;

[0027] Based on the user's recovery request, determine the scope and time point of the backup data to be recovered and select the most appropriate backup source for backup recovery.

[0028] Further, a method for performing a backup operation on ciphertext data includes:

[0029] Based on multi-point storage technology, perform local backup and off-site backup of the backup data;

[0030] Based on the automatic synchronization tool Dsynchronize, update the backup period in real time and perform integrity checks on the backup data to ensure that the backup data is not damaged and the data is consistent, and use the hash algorithm to verify the data integrity to ensure the consistency of the backup data at multiple storage locations.

[0031] Further, a method for performing backup recovery includes:

[0032] In response to the backup data being a full backup recovery: directly recover all data from the full backup file.

[0033] In response to the backup data being an incremental backup recovery: first recover the most recent full backup, and then sequentially apply the incremental backup files until the specified time point is restored;

[0034] In response to the backup data being a differential backup recovery: first recover the most recent full backup, and then apply the latest differential backup file to restore to the specified time point;

[0035] After the recovery is completed, verify the integrity and consistency of the backup data to ensure that the recovered data is available and accurate;

[0036] Confirm to the user that the data recovery is successful and record the recovery log for auditing and tracking.

[0037] Further, a method for evaluating and predicting the health status of users includes:

[0038] Based on the decrypted data, identify the sources of multi-modal data; among them, the multi-modal data types include microcirculation images of the nail fold, plantar and retina, numerical data of microcirculation blood flow, and text data of medical records;

[0039] Based on the image data, numerical data, and text data in the multi-modal data, perform single-modal feature extraction to obtain feature information, quantization metrics, and association information;

[0040] Based on feature-level fusion, integrate the single-modal features of feature information, quantization metrics, and association information to obtain a fusion vector;

[0041] Construct a multi-modal data analysis model and train the multi-modal data analysis model;

[0042] Based on the trained multi-modal data analysis model, analyze and predict the fusion vector to obtain health status evaluation, risk prediction, and personalized health advice.

[0043] Another technical solution adopted by this application is: a secure and interoperable multi-site microcirculation data management system, including:

[0044] A data acquisition layer, which is used to connect to an external database to obtain users' medical data, and perform data storage and interface design;

[0045] A blockchain layer, which is used to store and encrypt data, and store encrypted ciphertext data through a private chain and a consortium chain;

[0046] A federated learning layer, which uses federated learning technology to perform joint modeling on multi-party data to obtain a multi-modal data analysis model; among them, modeling does not require centralized storage, only sharing model parameters to protect the data privacy of all parties;

[0047] An application layer, which is used to connect to external access devices, so as to provide health assessment results based on data analysis to help users understand their own health conditions.

[0048] The beneficial effects of this application are: this application realizes the standardized storage, sharing and efficient analysis of medical data. By integrating nailfold, plantar and retinal microcirculation images, dynamic blood flow monitoring data and medical text records, the system supports the unified management and in-depth analysis of multi-modal data, and improves the accuracy of health management and disease prediction. The system adopts hybrid encryption technology and blockchain storage to ensure the security of data transmission and storage. At the same time, it uses federated learning technology to achieve cross-institutional collaboration, promote data sharing and distributed modeling. Based on multi-modal fusion analysis, the system can generate personalized health assessments, risk predictions and management recommendations, and improve the user experience through intuitive visualization tools. Through continuous optimization and dynamic learning, the system not only provides strong support for the early diagnosis and precision medicine of diseases, but also builds an efficient and secure data infrastructure for health management institutions and clinical applications. Brief Description of the Drawings

[0049] Figure 1 is a schematic flowchart of an embodiment of the secure and interoperable multi-site microcirculation data management method of this application;

[0050] Figure 2 is Figure 1 a schematic flowchart of an embodiment of step S1 in

[0051] Figure 3 is Figure 1 a schematic flowchart of an embodiment of step S3 in

[0052] Figure 4 is Figure 1 a schematic flowchart of an embodiment of step S4 in

[0053] Figure 5 is Figure 1 a schematic flowchart of an embodiment of step S5 in

[0054] Figure 6 is a schematic structural diagram of an embodiment of a secure and interoperable multi-site microcirculation data management system of the present application. Detailed implementation manners

[0055] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments.

[0056] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, the present invention may be implemented in other ways different from those described herein. Therefore, the present invention is not limited by the limitations of the specific embodiments disclosed below.

[0057] Refer to Figure 1 , Figure 1 is a schematic flowchart of an embodiment of a secure and interoperable multi-site microcirculation data management method of the present application, which includes:

[0058] Step S1. Design a data model of a standard data structure based on the FHIR standard to obtain microcirculation data.

[0059] Among them, step S1 includes:

[0060] Step S11. Select and customize FHIR resources based on the core functions of the data structure.

[0061] Specifically, determine the functions that the data needs to support, clarify the core functions that the data structure should support, such as patient health information management, medical examination result recording, and diagnostic report generation, and at the same time identify the key data types and their uses. This lays the foundation for subsequent resource selection and customization.

[0062] Select and customize FHIR resources, and select appropriate FHIR resources according to functional requirements. For example, Patient is used to record patient information, Observation stores medical observation results, ImagingStudy is used to save image data, DiagnosticReport is used to generate diagnostic conclusions, Procedure describes medical operations, and BodySite identifies the collection site. If the standard resources cannot fully meet the requirements, they can be customized through the extension mechanism of FHIR.

[0063] Step S12. Define the relationships between FHIR resources based on the FHIR resources to obtain standard resources.

[0064] Specifically, to define the relationships between resources and ensure data logical integrity, the relationships between resources need to be clarified. For example, Observation references patient information (Patient) through Observation.subject; Observation identifies the collection site (BodySite) through Observation.bodySite; ImagingStudy associates with patient information through ImagingStudy.subject; DiagnosticReport references the observation results or imaging data through DiagnosticReport.result; Procedure associates with patients and medical staff through Procedure.subject and Procedure.performer.

[0065] Step S13. Based on the HAPI FHIR framework and standard data, perform data storage and interface design to obtain microcirculation data.

[0066] Specifically, create a data model and Profile, extend and constrain the standard resources through FHIRProfile. When defining the Profile, the attribute range of the resources can be constrained. For example, the status in Observation can be restricted to a specific set of status values to ensure data consistency. At the same time, certain fields can be specified as required fields, such as the name and gender of the Patient, to meet the requirements of medical data integrity. In addition, using the extension mechanism (Extension) of FHIR, custom attributes can be added to the resources. For example, a specific measurement device ID can be added to Observation, or the source and technical information of image acquisition can be extended for ImagingStudy.

[0067] Create sample FHIR resources. To verify the feasibility and compatibility of the design, multiple sample FHIR resources can be created to cover key scenarios. Describe the basic information of the patient through the Patient resource, including name, gender, date of birth, etc.; use the Observation resource to record medical observation results, clarifying the associated patient information and collection site; then store medical image data through the ImagingStudy resource, including the image source and relevant technical details; finally, use the DiagnosticReport resource to generate a diagnostic report and reference the relevant Observation and ImagingStudy resources.

[0068] In data storage and interface design, the HAPI FHIR framework is used to build the storage system, leveraging its powerful FHIR resource management capabilities to efficiently store and manage all standardized resources. The interface design is based on RESTful API, supporting create, read, update, and delete (CRUD) operations on FHIR resources. For example, information about a specific patient can be queried via GET / Patient / {id}, medical observation results can be uploaded using POST / Observation, or a specific diagnostic report can be retrieved via GET / DiagnosticReport / {id}. Additionally, the interface supports parameterized queries and filtering functions, such as filtering observation results by patient ID or a specified time range for testing and validation.

[0069] Documentation and Iteration: In the documentation and iterative optimization phase, it is necessary to document in detail the data model, extensions, API interfaces, and test cases to ensure good traceability of the design process and implementation details. This includes the definition of each FHIR resource, the specific implementation of extensions, and the description of interface functions, while also recording test cases and verification results to provide clear guidance for development and maintenance. In iterative optimization, by collecting user feedback, analyzing problems and requirements in actual use, the data model and interface design are dynamically adjusted.

[0070] Step S2. Preprocess the microcirculation data to obtain preprocessed data.

[0071] Specifically, perform preliminary processing such as data denoising and standardization on the microcirculation data to ensure data quality, and finally output the data to generate high-quality preprocessed data.

[0072] Step S3. Based on hybrid encryption technology, encrypt the transmission and storage processes of the preprocessed data to obtain ciphertext data.

[0073] Among them, step S3 includes:

[0074] Step S31. Generate a random symmetric first encryption key.

[0075] Specifically, when the preprocessed data is input, the system generates a random symmetric first encryption key (AES key).

[0076] Step S32. Based on an asymmetric encryption key, encrypt the first encryption key to obtain a second encryption key.

[0077] Specifically, use asymmetric encryption (RSA) to encrypt the first encryption key: use the recipient's RSA public key to encrypt the generated AES symmetric key, and the encrypted AES key is called the second encryption key.

[0078] Step S33. Encrypt the preprocessed data based on the second encryption key to obtain ciphertext data.

[0079] Specifically, encrypt the data using the second encryption key: encrypt the preprocessed data using the generated AES symmetric key to generate ciphertext data.

[0080] Step S34. Store the second encryption key and the ciphertext data in the blockchain, generate a hash fingerprint of the ciphertext data and store it in the blockchain to ensure the immutability of the data.

[0081] Specifically, transmit the encrypted AES key and the ciphertext data together to the recipient or storage system. The data transmission is carried out through the secure protocol TLS to ensure security during the transmission process.

[0082] Store the ciphertext data in the blockchain layer (private chain or consortium chain). In the blockchain layer, store the hash value of the ciphertext data and the smart contract record to ensure the integrity and immutability of the data.

[0083] Control the access conditions and permissions of the data through smart contract records. Generate a hash fingerprint of the data and store it in the blockchain to ensure the immutability of the data.

[0084] When an authorized user requests access to the data, the system verifies the user's permissions through the smart contract. If the permission verification passes, the system decrypts the encrypted AES key using the RSA private key. Use the decrypted AES key to decrypt the ciphertext data stored in the blockchain layer, recover the original data, and finally provide the decrypted data to the authorized user.

[0085] Step S4. Based on the multi-level backup and recovery technology, perform data backup and recovery on the ciphertext data to obtain decrypted data.

[0086] Among them, step S4 includes:

[0087] Step S41. Evaluate the importance of the ciphertext data to determine the backup strategy combination; among them, the backup strategy combination includes full backup, incremental backup, and differential backup.

[0088] Specifically, analyze the importance, change frequency, and storage requirements of the data to determine the applicable backup strategy combination, and then select the appropriate backup strategy. Among them, the backup strategy combination includes full backup, incremental backup, and differential backup.

[0089] Full backup: Suitable for critical data or low-frequency backup requirements to ensure the integrity of the data.

[0090] Incremental backup: Suitable for high-frequency backup requirements, only back up the data that has changed since the last backup, saving storage space and backup time.

[0091] Differential backup: It is between full backup and incremental backup, backing up the data that has changed since the last full backup, combining the advantages of both.

[0092] Step S42. Based on the backup policy combination, perform a backup operation on the ciphertext data to obtain backup data.

[0093] Specifically, adopt multi-location storage technology for local backup or off-site backup. Among them, local backup means storing the backup data in a local storage device to ensure fast access and recovery. Off-site backup means storing the backup data in multiple geographically separated nodes to prevent data loss due to natural disasters or other emergencies.

[0094] At the same time, use the automatic synchronization tool Dsynchronize to update in real time, perform an integrity check on the backup data to ensure that the backup file is not damaged and the data is consistent, and use the hash algorithm to verify the data integrity to ensure the data consistency of multiple storage locations.

[0095] Step S43. Regularly perform a recovery test on the backup data, simulate the data recovery process, and verify the availability of the backup for the backup data and the effectiveness of the recovery process.

[0096] Specifically, perform a recovery test regularly, simulate the data recovery process, verify the availability of the backup data and the effectiveness of the recovery process. Record the test results, identify and fix potential problems in the recovery process.

[0097] Step S44. Based on the user's recovery request, determine the scope and time point of the backup data to be recovered and select the most suitable backup source for backup recovery.

[0098] Specifically, receive a recovery request from the user or the system, determine the data range and time point to be recovered, and select the most suitable backup source (full backup, incremental backup or differential backup) for backup recovery, as follows:

[0099] Full backup recovery: Directly recover all data from the full backup file.

[0100] Incremental backup recovery: First recover the most recent full backup, and then sequentially apply the incremental backup files until the specified time point is reached.

[0101] Differential backup recovery: First recover the most recent full backup, and then apply the latest differential backup file to recover to the specified time point.

[0102] After the recovery is completed, verify the integrity and consistency of the data to ensure that the recovered data is available and accurate.

[0103] Finally, confirm the successful data recovery to the user or the system, and record the recovery log for auditing and tracing.

[0104] Step S5. Based on the multimodal prediction model and the decrypted data, evaluate and predict the user's health status.

[0105] Among them, step S5 includes:

[0106] Step S51. Based on the decrypted data, identify the sources of multimodal data; among them, the types of multimodal data include microcirculation images of three parts: nail fold, plantar and retina, numerical data of microcirculation blood flow, and text data of medical records.

[0107] Specifically, to identify the sources of multimodal data, the types of multimodal data to be processed include microcirculation images of three parts: nail fold, plantar and retina, numerical data of microcirculation blood flow, and text data of medical records. These data are standardized and represented through the FHIR standard to provide support for subsequent analysis and system interoperability.

[0108] Step S52. Based on the image data, numerical data, and text data in the multimodal data, perform single-modal feature extraction to obtain feature information, quantization metrics, and association information.

[0109] Specifically, for the image data, it is necessary to extract its deep-level feature information to capture the changes in the morphology and function of microvessels; for the numerical data, extract its key quantization metrics; for the text data, extract the association information from the semantic level. The extraction of these single-modal features provides a high-quality basis for subsequent multimodal fusion and analysis.

[0110] Step S53. Based on feature-level fusion, integrate the single-modal features of feature information, quantization metrics, and association information to obtain a fusion vector.

[0111] Specifically, adopt the method of feature-level fusion. Through the method of direct splicing, integrate the single-modal features of images, numerical values, and texts into a unified fusion vector. This method is simple and efficient, and can achieve the integration of cross-modal information while ensuring the integrity of each modal feature, providing the model with rich feature expression capabilities.

[0112] Step S54. Build a multimodal data analysis model and train the multimodal data analysis model.

[0113] Specifically, when building a multimodal data analysis model, due to its powerful feature interaction modeling ability, it can efficiently capture the complex relationships between multimodal features. In the model training stage, an optimization method is adopted to improve the performance of the model. When training the model, the dataset is divided into a training set, a validation set, and a test set to ensure that the model has good generalization ability.

[0114] Step S55. Based on the trained multi-modal data analysis model, analyze and predict the fusion vector to obtain health status assessment, risk prediction, and personalized health advice.

[0115] Specifically, after training, the model is converted into a deployment format adapted to the actual environment and integrated into the health management system. This system supports real-time data input and prediction output, providing users with health status assessment, risk prediction, and personalized health advice, thus playing an important role in daily health management and clinical applications.

[0116] In practical applications, by continuously collecting new data and expanding the dataset, the analysis model is dynamically updated to adapt to changes in data distribution. At the same time, the model performance is regularly evaluated, and the feature selection and data fusion methods are optimized to ensure the long-term effectiveness and accuracy of the system. This process enables the health management system to continuously improve service quality and provide users with more accurate health advice and prediction support.

[0117] Refer to Figure 6 , Figure 6 FIG. is a schematic structural diagram of an embodiment of the safe and interoperable multi-site microcirculation data management system of the present application. The system includes: a data acquisition layer 1, a blockchain layer 2, a federated learning layer 3, and an application layer 4.

[0118] The data acquisition layer 1 is used to connect to an external database to obtain users' medical data, perform data storage and interface design, and perform data preprocessing to obtain preprocessed data.

[0119] Among them, at the grass-roots level of the hospital or in the health management center, the process of the data acquisition layer 1 collecting patients' health data aims to efficiently and accurately capture microcirculation data and related health information. First, a variety of medical devices and sensors are deployed at the collection site, including microcirculation microscopes, dynamic blood flow monitors, and image capture devices, to meet the collection needs of different parts and data types.

[0120] The above devices are interconnected through a unified system integration platform, supporting real-time data input and multi-modal data management.

[0121] After the patient arrives, first establish or update an electronic health record for him / her, record basic information including name, gender, age, and medical history, and provide a unique identifier for subsequent data management.

[0122] Subsequently, the technician calibrates the devices to ensure the accuracy and consistency of the collection, and guides the patient to place the correct posture according to the collection requirements, place the finger under the microscope for nail fold collection, place the sole on a dedicated scanner for scanning, and collect retinal images through a fundus imaging device.

[0123] During the actual acquisition process, the microcirculation microscope device captures high-resolution images of the nail fold, plantar surface, and retina in real time, and the dynamic blood flow monitor records the velocity, flow rate, and other relevant parameters of the patient's blood flow.

[0124] Perform multi-source data integration: Ensure the consistency and integrity of the data, and integrate different types and formats of health data. Suppose the nail fold image data of a patient is collected by a microcirculation microscope and saved as a high-resolution TIFF file, the dynamic blood flow data is exported in the form of a CSV file, and the medical history data provided by the electronic health record (EHR) system is unstructured text content. Through the data integration platform, the image data will be converted into the DICOM format, the blood flow data will be filled into the FHIR Observation resource through a data mapping tool, and the important diagnostic information will be extracted from the medical history text after semantic analysis and filled into the FHIR Condition resource. Finally, these multi-source data are integrated into the same electronic health record, bound to the patient's unique ID, and form a complete health record.

[0125] Blockchain layer 2 is used to store and encrypt data, and stores encrypted ciphertext data through private chains and consortium chains. Among them, blockchain layer 2 stores and encrypts data, and stores encrypted health data through private chains and consortium chains; uses hybrid encryption technology (a combination of symmetric encryption and asymmetric encryption) to ensure the security and efficiency of data; uses the asymmetric encryption algorithm RSA to encrypt and transmit the symmetric encryption key, and then uses the symmetric encryption algorithm AES to encrypt the specific data.

[0126] Blockchain layer 2 can control permissions, uses smart contracts to define data access conditions, and ensures that only authorized users (doctors, patients, and primary hospitals) can access specific data; the access records of blockchain layer 2 are traceable to ensure transparency. Blockchain layer 2 can protect data privacy, restrict data access rights through private chains, and only allow specific institutions or users to participate.

[0127] The consortium chain supports multi-party collaboration, but still strictly restricts data access. Tamper-proof and traceability, the immutability of blockchain technology ensures the authenticity of data; every storage and access of data is recorded on the chain and can be traced at any time.

[0128] The federated learning layer 3 uses federated learning technology to perform joint modeling on multi-party data to obtain a multi-modal data analysis model; among them, modeling does not require centralized storage, only needs to share model parameters, and protects the data privacy of all parties. The federated learning layer 3 is applied to complex medical data analysis through intelligent analysis, including disease prediction, treatment effect evaluation, and health risk management. Utilize multi-source heterogeneous data to improve the accuracy and generalization ability of the model.

[0129] The application layer 4 is used to connect external access devices, so as to provide health assessment results based on data analysis, helping users understand their own health conditions. Interactive health analysis provides health assessment results based on data analysis to help users understand their own health conditions; the application layer 4 provides personalized health risk prediction and advice. Intuitive charts such as bar charts and dashboards are used to display the dynamic changes of health data, reflecting the changes in microcirculation images and the historical trends of monitoring indicators, realizing visual display; 3D models and dynamic demonstrations are supported to facilitate communication between doctors and patients. The personalized health management of the application layer 4 provides a customized health management plan according to the user's health data, including diet, exercise and medication use advice; health reminders are provided, including medication taking and physical examination appointments.

[0130] Multi-device access allows multiple terminals such as mobile phones, tablets, and computers to access, facilitating users to view data anytime and anywhere. Third-party interfaces are supported, compatible with the integration of electronic health record (EHR) systems, insurance platforms, and health applications; support for providing anonymous data interfaces for research institutions to promote medical research.

[0131] This application aims to achieve the standardized storage, sharing, and efficient analysis of medical data. By integrating nailfold, plantar, and retinal microcirculation images, dynamic blood flow monitoring data, and medical text records, the system supports the unified management and in-depth analysis of multi-modal data, improving the accuracy of health management and disease prediction. The system uses hybrid encryption technology and blockchain storage to ensure the security of data transmission and storage. At the same time, federated learning technology is used to achieve cross-institutional collaboration, promoting data sharing and distributed modeling. Based on multi-modal fusion analysis, the system can generate personalized health assessments, risk predictions, and management advice, and improve the user experience through intuitive visualization tools. Through continuous optimization and dynamic learning, this system not only provides strong support for the early diagnosis and precision medicine of diseases, but also builds an efficient and secure data infrastructure for health management institutions and clinical applications.

[0132] The above are only the embodiments of this application, and do not limit the patent scope of this application. Any equivalent structure or equivalent process transformation made using the content of the specification and drawings of this application, or directly or indirectly applied in other related technical fields, shall be included in the patent protection scope of this application by the same token.

Claims

1. A secure and interoperable multi-site microcirculation data management method, characterized in that, Including: Based on the FHIR standard, design a data model with a standard data structure to obtain microcirculation data; Preprocess the above-mentioned microcirculation data to obtain preprocessed data; Based on hybrid encryption technology, encrypt the transmission and storage processes of the preprocessed data to obtain ciphertext data; Based on multi-level backup and recovery technology, perform data backup and recovery on the ciphertext data to obtain decrypted data; Based on a multi-modal prediction model and the decrypted data, evaluate and predict the user's health status.

2. The method according to claim 1, characterized in that, The method for designing the data structure includes: Select and customize FHIR resources based on the core functions of the data structure; Based on the FHIR resources, define the relationships between the FHIR resources to obtain standard resources; Based on the HAPI FHIR framework and the standard data, perform data storage and interface design to obtain the microcirculation data.

3. The method according to claim 1, wherein The method for obtaining ciphertext data includes: Generate a random symmetric first encryption key; Based on an asymmetric encryption key, encrypt the first encryption key to obtain a second encryption key; Based on the second encryption key, encrypt the preprocessed data to obtain ciphertext data; Store the second encryption key and the ciphertext data in the blockchain, generate a hash fingerprint of the ciphertext digital data and store it in the blockchain to ensure the immutability of the data.

4. The method according to claim 1, characterized in that, The method for performing data backup and recovery on the ciphertext data includes: Evaluate the importance of the ciphertext data to determine a backup strategy combination; wherein, the backup strategy combination includes full backup, incremental backup, and differential backup; Based on the backup strategy combination, perform a backup operation on the ciphertext data to obtain backup data; Regularly perform a recovery test on the backup data, simulate the data recovery process, and verify the availability of the backup for the backup data and the effectiveness of the recovery process; Based on the user's recovery request, determine the scope and time point of the backup data to be recovered and select the most appropriate backup source for backup recovery.

5. The method according to claim 4, characterized in that The method for performing a backup operation on the ciphertext data includes: Based on multi-point storage technology, perform local backup and off-site backup on the backup data; Based on the automatic synchronization tool Dsynchronize, update the backup period in real time and perform an integrity check on the backup data to ensure that the backup data is not damaged and the data is consistent, and use a hash algorithm to verify the data integrity to ensure the consistency of the backup data at multiple storage locations.

6. The method according to claim 4, wherein The method for performing backup recovery includes: In response to the backup data being a full backup recovery: directly recover all data from the full backup file. In response to the backup data being an incremental backup recovery: first recover the most recent full backup, and then sequentially apply the incremental backup files until the specified time point is reached; In response to the backup data being a differential backup recovery: first recover the most recent full backup, and then apply the latest differential backup file to recover to the specified time point; After the recovery is completed, verify the integrity and consistency of the backup data to ensure that the recovered data is available and accurate; Confirm to the user that the data recovery is successful and record the recovery log for auditing and tracking.

7. The method according to claim 1, characterized in that, A method for evaluating and predicting the health status of users, including: Based on the decrypted data, identifying the sources of multimodal data; wherein, the types of the multimodal data include microcirculation images of three parts, namely nail folds, plantar surfaces, and retinas, numerical data of microcirculation blood flow, and text data of medical records; Based on the image data, numerical data, and text data in the multimodal data, performing single-modal feature extraction to obtain feature information, quantization indexes, and association information; Based on feature-level fusion, integrating the single-modal features of the feature information, quantization indexes, and association information to obtain a fusion vector; Constructing a multimodal data analysis model and training the multimodal data analysis model; Based on the trained multimodal data analysis model, analyzing and predicting the fusion vector to obtain health status evaluation, risk prediction, and personalized health advice.

8. A secure and interoperable multi-site microcirculation data management system, characterized in that, Including: A data collection layer for connecting to an external database to obtain users' medical data and performing data storage and interface design; A blockchain layer for storing and encrypting data, and storing the encrypted ciphertext data through a private chain and a consortium chain; A federated learning layer for jointly modeling on multi-party data using federated learning technology to obtain a multimodal data analysis model; wherein, the modeling does not require centralized storage, only sharing model parameters to protect the data privacy of all parties; An application layer for connecting to an external access device to provide health assessment results based on data analysis to help users understand their own health status.

Images