Cardiovascular monitoring system based on artificial intelligence

By integrating multiple terminal devices and artificial intelligence technologies, multimodal data fusion and personalized health management of cardiovascular monitoring systems are solved, and the problems of monitoring discontinuity and data privacy leakage in the existing technology are improved, and the accuracy of monitoring is improved. and safety, reducing the mortality and disability rates of cardiovascular disease.

CN118969309APending Publication Date: 2024-11-15QINGDAO HAINAN HEALTH CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411022355.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing cardiovascular monitoring systems lack the ability to integrate multimodal data, and have problems such as discontinuity of monitoring, data privacy leakage, low accuracy, low user acceptance, and high cost.

Method used

Design a cardiovascular monitoring system based on artificial intelligence, integrates multiple terminal devices, adopts multi-modal data fusion unit, artificial intelligence analysis unit, privacy protection unit and emergency processing module, through data collection, processing, storage, fusion and analysis , generate comprehensive feature vectors to achieve accurate monitoring and early warning of cardiovascular diseases, and provide personalized health management and emergency response solutions.

Benefits of technology

It realizes comprehensive and continuous monitoring of cardiovascular health, improves the accuracy and security of monitoring, ensures data privacy, reduces the mortality and disability rate of cardiovascular diseases, and provides a cost-effective health management solution.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118969309A_ABST
    Figure CN118969309A_ABST
Patent Text Reader

Abstract

The invention discloses a cardiovascular monitoring system based on artificial intelligence, and relates to the technical field of medical monitoring. And data acquisition, processing, storage, fusion, analysis, privacy protection and early warning functions are realized through the terminal equipment and the central processing system. The system collects various data such as electrocardiogram and cardiac magnetic resonance imaging, and the data is classified and subjected to strategy processing through the data processing module and stored in public and private databases. The multi-modal data fusion unit generates a comprehensive feature vector, and the artificial intelligence analysis unit realizes disease monitoring and early warning. The privacy protection unit ensures data security, and the display and early warning module graphically displays an analysis result and gives an early warning in case of an abnormal condition. The emergency processing module provides an emergency processing scheme for the early warning signal. The system comprehensively and efficiently serves the management, prevention and control of cardiovascular diseases, and has the advantages of higher accuracy, better privacy protection, higher adaptability, lower cost and higher transparency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of medical monitoring, and in particular relates to a cardiovascular monitoring system based on artificial intelligence. Background Art

[0002] With the increasing trend of population aging and changes in lifestyle, cardiovascular disease has become one of the leading causes of death and disability worldwide. Traditional cardiovascular disease monitoring methods mainly rely on professional equipment in hospitals, which have problems such as discontinuous monitoring, poor patient experience, and uneven distribution of medical resources. In recent years, with the rapid development of technologies such as the Internet of Things, big data, and artificial intelligence, cardiovascular monitoring systems based on remote monitoring and intelligent analysis have gradually become a research hotspot.

[0003] Most existing cardiovascular monitoring systems focus on the collection and analysis of a single data source, such as using only electrocardiogram (ECG) or blood pressure monitoring data, lacking the ability to fuse multimodal data, making it difficult to fully and accurately reflect the health status of patients. In addition, there are many other defects, such as data privacy leakage, low monitoring accuracy, low user acceptance, high cost, and lack of model transparency. In order to solve these problems, it is necessary to develop a new cardiovascular monitoring system. Summary of the invention

[0004] The purpose of the present invention is to provide a cardiovascular monitoring system based on artificial intelligence to solve the existing problems mentioned in the background technology.

[0005] In order to solve the above technical problems, the present invention is achieved through the following technical solutions:

[0006] The present invention is an artificial intelligence-based cardiovascular monitoring system, comprising a terminal device and a central processing system for data transmission with the terminal device, wherein the central processing system comprises:

[0007] Data collection module: collects patients' rehabilitation-related data from various terminal devices and generates data sets. The collected data include but are not limited to electrocardiogram collection, cardiac magnetic resonance imaging collection, cardiac auscultation collection, photoplethysmography collection, blood pressure monitoring and blood biochemical index collection;

[0008] Data processing module: classifies the data sets obtained by different collection methods and selects specific processing strategies according to different data set types;

[0009] Data storage module: partitions and stores the data sets processed by the data processing module, and collects the data into the public pathology database and the patient private database;

[0010] Multimodal data fusion unit: uses a fusion model to fuse preprocessed data sets of different modalities to generate a comprehensive feature vector;

[0011] Artificial intelligence analysis unit: uses pre-trained adaptive learning algorithms to analyze comprehensive feature vectors to achieve cardiovascular disease monitoring and early warning;

[0012] Privacy protection unit: encrypts and de-identifies data to ensure data security;

[0013] Display and warning module: displays the analysis results to the user in a graphical interface and issues warnings when abnormal conditions are detected;

[0014] Emergency processing module: classifies warning signals and provides emergency processing plans when the warning signals exceed the set threshold.

[0015] As a preferred technical solution of the present invention, the terminal device is a fixed monitoring device of a medical institution or a wearable device that can be conveniently used by patients. The terminal device is provided with a data acquisition unit, a data output and receiving unit, an early warning prompt unit and a user information management unit.

[0016] As a preferred technical solution of the present invention, the user information management unit stores the user's personal identity information, the user's historical health data set, and the future model of the disease generated by the medical model and continuously adjusted according to the user's health data. The user management unit can adjust the user's identity information, emergency contacts, health data, and warning thresholds.

[0017] As a preferred technical solution of the present invention, when the terminal device is a wearable device, the terminal device is installed with a lightweight algorithm data processing system; at the same time, the terminal device also includes a telemedicine service module, which automatically matches the medical institution with corresponding processing authority based on the warning level and emergency treatment plan and conducts active calls and medical consultations.

[0018] As a preferred technical solution of the present invention, the data processing module classifies the data set according to the data type and source, adopts denoising, normalization and feature extraction methods when processing time series data; adopts image enhancement, separation and feature extraction methods when processing image data; and adopts data cleaning and data conversion methods when processing structured data.

[0019] As a preferred technical solution of the present invention, the data processing module adopts parallel processing or stream processing or a combination of batch processing and real-time processing when processing multiple data sets.

[0020] As a preferred technical solution of the present invention, when parallel processing is adopted, the MapReduce model is adopted; when batch processing is combined with real-time processing, the Lambda architecture is adopted; when stream processing is adopted, Apache Kafka Streams or Apache Flink is adopted.

[0021] As a preferred technical solution of the present invention, the difference between the patient private database and the public pathology database in the data storage module is that the data sets in the patient private database are matched one by one with the corresponding patient identity information, and an independent encrypted micro private database is matched one by one in the patient private database according to the patient identity information.

[0022] As a preferred technical solution of the present invention, the fusion model is a deep belief network, and the deep belief network is formed by stacking multiple restricted Boltzmann machines (RBMs).

[0023] As a preferred technical solution of the present invention, the display and warning module includes a threshold setting submodule, which sets individualized physiological parameter thresholds in combination with statistical analysis methods, and generates parameter threshold standards through a threshold adjustment mechanism.

[0024] The present invention has the following beneficial effects:

[0025] This system achieves comprehensive monitoring of the patient's cardiovascular health status by integrating a variety of terminal devices, such as electrocardiographs, MRI scanners, and sphygmomanometers. This not only improves the continuity and timeliness of monitoring, but also provides us with a rich data source to support subsequent analysis. The system uses a multimodal data fusion unit that can integrate data from different modalities, such as electrocardiograms, MRI images, auscultation, etc., to generate a comprehensive feature vector. This fusion method can more comprehensively and accurately reflect the patient's health status and improve the accuracy and reliability of monitoring. The system can dynamically adjust the threshold according to the patient's specific health status and monitoring data to generate personalized physiological parameter threshold standards. This provides patients with a more accurate health management plan and helps to detect potential cardiovascular disease risks at an early stage. The system also uses encryption and de-identification technology to ensure the security of data during transmission, storage, and processing. At the same time, the data is divided into a public pathology database and a patient private database, which not only promotes the progress of medical research, but also ensures the security and privacy protection of patients' personal information. The system has an emergency processing module that can provide emergency emergency processing plans according to the severity of the warning signal. At the same time, the telemedicine service module can automatically match the corresponding medical institutions according to the warning level, provide patients with immediate medical assistance, and effectively reduce the mortality and disability rates of cardiovascular diseases.

[0026] In general, this AI-based cardiovascular monitoring system integrates multiple technologies and methods to achieve the advantages of continuous monitoring, multimodal data fusion, personalized health management, privacy protection, emergency response, low cost and high efficiency. It provides strong support for the prevention, monitoring and treatment of cardiovascular diseases, and brings significant economic benefits to patients and medical institutions.

[0027] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0029] Figure 1 It is a schematic diagram of the architecture of the present invention;

[0030] Figure 2 A schematic diagram of the internal system of the terminal device when the terminal device is a wearable mobile device in the present invention;

[0031] Figure 3 It is a working schematic diagram of the fusion algorithm in the present invention;

[0032] Figure 4 Schematic diagram of the working method of the threshold setting submodule in the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] like Figure 1As shown: The present invention provides a cardiovascular monitoring system based on artificial intelligence, including a terminal device and a central processing system for transmitting data with the terminal device, the central processing system including a data acquisition module: collecting rehabilitation-related data of patients from various terminal devices and generating a data set, the data collected includes but is not limited to electrocardiogram collection, cardiac magnetic resonance imaging collection, cardiac auscultation collection, photoplethysmography collection, blood pressure monitoring and blood biochemical index collection; a data processing module: classifying the data sets obtained by different collection methods and selecting specific processing strategies according to different data set types; a data storage module: classifying the data sets processed by the data processing module The data is stored in the same area and aggregated into the public pathology database and the patient's private database; multimodal data fusion unit: the fusion model is used to fuse the pre-processed data sets of different modalities to generate a comprehensive feature vector; artificial intelligence analysis unit: the pre-trained adaptive learning algorithm is used to analyze the comprehensive feature vector to achieve cardiovascular disease monitoring and early warning; privacy protection unit: the data is encrypted and de-identified to ensure data security; display and early warning module: the analysis results are displayed to the user in a graphical interface, and an early warning is issued when an abnormal situation is detected; emergency processing module: the early warning signal is classified, and when the early warning signal exceeds the set threshold, an emergency response plan is provided.

[0035] In this embodiment, the system of the present invention includes two sections, one is a central processing system for processing data, and the other is a terminal device for specific monitoring. The central processing system includes the following units and modules:

[0036] Data collection module, which is responsible for collecting patients' cardiovascular health-related data from a variety of terminal devices. These devices include but are not limited to electrocardiographs, MRI scanners, stethoscopes, pulse oximeters, sphygmomanometers, and laboratory biochemical analyzers. By integrating multiple data collection methods, more comprehensive and detailed patient health information can be obtained, providing a rich data source for subsequent analysis. Comprehensive data collection improves the accuracy and comprehensiveness of the monitoring system, helps to detect potential cardiovascular disease risks at an early stage, and provides data support for personalized treatment plans.

[0037] Data processing module: This module automatically selects and applies the corresponding preprocessing algorithm according to different data types (such as ECG waveforms, image data, audio recordings, etc.). The processing steps may include noise filtering, signal enhancement, feature extraction, etc. to ensure data quality and extract valuable information for subsequent analysis. Through intelligent data processing, invalid and redundant information is reduced, the efficiency and accuracy of data analysis are improved, and a solid foundation is laid for subsequent data fusion and disease diagnosis.

[0038] The data storage module is responsible for storing the processed data securely in the cloud or on a local server. The processed data is partitioned and stored in two main databases: a public pathology database and a patient private database. The public pathology database is accessible to researchers and medical professionals to promote cardiovascular disease research and knowledge sharing. The public pathology database can also be used for scientific research, algorithm optimization and other purposes. The patient private database strictly abides by privacy protection regulations to ensure that only authorized personnel can access patients' sensitive information. This partitioned storage strategy not only promotes the advancement of medical research, but also ensures the security and privacy protection of patients' personal information, in line with the ethical and legal requirements of the medical industry.

[0039] The multimodal data fusion unit uses advanced fusion algorithms to integrate data sets from different modalities (such as images, sounds, physiological signals, etc.) into a comprehensive feature vector. This process takes into account the complementarity and correlation between the data of each modality, aiming to extract more comprehensive and representative information on the patient's health status. Multimodal data fusion significantly improves the system's ability to identify and predict cardiovascular diseases, making precision medicine possible.

[0040] The artificial intelligence analysis unit is based on advanced technologies such as deep learning and machine learning. This unit uses a pre-trained adaptive learning algorithm to conduct in-depth analysis of the comprehensive feature vector to achieve intelligent monitoring and early warning of cardiovascular diseases. The algorithm can automatically identify abnormal patterns and compare them with known disease characteristics to make preliminary diagnoses or early warnings. At the same time, the adaptive learning algorithm can continuously adjust its prediction model based on new data to improve the accuracy of monitoring and early warning. This means that the system can become smarter over time and better adapt to changes in individual patients. Intelligent analysis greatly improves the efficiency and accuracy of disease diagnosis, reduces human errors, and wins valuable treatment time for patients.

[0041] The privacy protection unit uses encryption technology and de-identification methods to ensure data security during data transmission, storage and processing. The unit is also responsible for monitoring data access rights to prevent unauthorized access and data leakage. It strengthens the data protection mechanism, enhances patients' trust in medical services, and provides security for the long-term stable operation of the system.

[0042] The display and warning module displays the analysis results to users in an intuitive and easy-to-understand graphical interface, such as heart rate trend chart, blood pressure change curve, etc. At the same time, when an abnormal situation is detected, a warning is issued in a timely manner through various means such as sound, light flash, and mobile phone notification. Intuitive display and timely warning help patients and medical staff quickly understand the changes in their condition and take corresponding measures, effectively reducing the incidence of acute events of cardiovascular diseases.

[0043] The emergency treatment module is the last line of defense of the system. When the warning signal exceeds the preset threshold, the module will provide a series of emergency response measures, grade the warning signal according to its severity, and provide corresponding emergency response plans. The plan may include emergency medical guidance, first aid knowledge push, automatic contact with the emergency center, etc. The emergency treatment module provides fast and effective response measures in emergency situations, effectively reducing the mortality and disability rates of cardiovascular diseases and ensuring the life safety of patients.

[0044] In another optional embodiment, the terminal device is a fixed monitoring device of a medical institution or a wearable device that the patient can carry with him for easy use. The terminal device is provided with a data acquisition unit, a data output and receiving unit, an early warning prompt unit and a user information management unit. The user information management unit stores the user's personal identity information, the user's historical health data set, and the future model of the disease condition generated by the medical model and continuously adjusted according to the user's health data. The user management unit can adjust the user's identity information, emergency contacts, health data and early warning thresholds. When the terminal device is a wearable device, the terminal device is equipped with a lightweight algorithm data processing system; at the same time, the terminal device also includes a telemedicine service module, which automatically matches the medical institution with the corresponding processing authority based on the early warning level and the emergency treatment plan, and makes active calls and medical consultations.

[0045] In this embodiment, the design and function of the terminal device make it an indispensable part of the patient's daily health management. Whether it is a fixed monitoring device in a medical institution or a convenient wearable device that patients can carry with them, they are equipped with advanced data acquisition units that can collect key physiological parameters in real time.

[0046] The data output and receiving unit ensures the smooth exchange of information. It can not only send the collected data to the central processing system, but also receive instructions and feedback from the central system. This two-way communication ensures the real-time and interactive nature of the monitoring system, allowing patients and medical workers to communicate and intervene in a timely manner.

[0047] The early warning unit is an important part of ensuring patient safety. It sends visual or auditory signals based on the analysis results to alert patients and medical staff to potential health risks. For example, if an abnormal heart rhythm is detected, the wearable device will alert the user by vibrating or emitting an alarm sound, while providing a brief description of the situation through the display.

[0048] The user information management unit provides a comprehensive health profile for each user. It not only stores the user's personal identity information, but also includes historical health data sets and personalized future models of disease conditions. These models are generated by medical algorithms and dynamically adjusted based on the user's real-time data to predict future health conditions and take preventive measures in advance.

[0049] Especially when the terminal device is a wearable device, lightweight algorithm data processing systems are particularly important. Due to the limited computing resources of wearable devices, these lightweight algorithms are optimized to quickly process data and give feedback without sacrificing accuracy. This allows users to perform preliminary health monitoring and assessment on the spot even without a network connection.

[0050] The remote medical service module further expands the functionality of the terminal device. It can automatically search and establish contact with appropriate medical institutions based on the warning level and emergency treatment plan. In emergency situations, such as when an acute cardiac event is detected, the module can immediately initiate a call service to connect the patient with a medical expert with corresponding treatment capabilities and provide real-time medical consultation and support.

[0051] This intelligent and automated telemedicine service not only provides patients with timely medical intervention, but also greatly reduces the pressure on the emergency medical system and optimizes the allocation and use of resources. For patients, this means that they can get professional and immediate medical assistance no matter where they are.

[0052] In another optional embodiment, the data processing module classifies the data set according to the data type and source, and uses denoising, normalization and feature extraction methods when processing time series data; uses image enhancement, separation and feature extraction methods when processing image data; and uses data cleaning and data conversion methods when processing structured data.

[0053] The data processing module uses parallel processing or stream processing or a combination of batch processing and real-time processing when processing multiple data sets. When parallel processing is used, the MapReduce model is used; when batch processing is combined with real-time processing, the Lambda architecture is used; when stream processing is used, Apache Kafka Streams or Apache Flink is used.

[0054] In this embodiment, the data processing module efficiently processes different types of data sets in the above manner to ensure the real-time and accuracy of data processing. Specifically, for time series data, the data processing module first performs denoising and normalization processing to remove noise and make the data comparable. Then, through feature extraction techniques such as time domain analysis, frequency domain analysis, etc., key features useful for cardiovascular monitoring are extracted. For image data, the data processing module first performs image enhancement and segmentation processing to improve the clarity and readability of the image. Then, through feature extraction techniques such as color histograms, texture features, etc., key features useful for cardiovascular monitoring are extracted. For structured data, the data processing module first performs data cleaning and conversion processing to remove invalid or erroneous data and convert the data into a unified format. Then, through feature extraction techniques such as principal component analysis (PCA), linear discriminant analysis (LDA), etc., key features useful for cardiovascular monitoring are extracted.

[0055] At the same time, the data processing module can use different processing methods to efficiently process multiple data sets, including parallel processing, stream processing, and a combination of batch processing and real-time processing. The following is a detailed description of the model and calculation method of each processing method:

[0056] 1. Parallel processing: using the MapReduce model

[0057] MapReduce model: includes:

[0058] 1.1, Map phase:

[0059] Input: The segmented dataset.

[0060] Map: Apply the Map function to each data block to generate an intermediate result of key-value pairs.

[0061] Output: Intermediate key-value pairs, stored in the local file system.

[0062] 1.2 Reduce phase:

[0063] Input: intermediate key-value pairs.

[0064] Reduction: Apply the Reduce function to all values ​​for each key to produce the final result.

[0065] Output: The final result, usually a list of key-value pairs.

[0066] The calculation method using the MapReduce model is:

[0067] Data segmentation: Split the data set into multiple data blocks, each of which is processed by a Map task.

[0068] Task allocation: The MapReduce framework is responsible for allocating data blocks to different Map tasks and executing these tasks in parallel on multiple computing nodes;

[0069] Intermediate result storage: The intermediate key-value pairs generated by the Map task are stored in the local file system, waiting to be read by the Reduce task;

[0070] Data locality: Reduce tasks will try to select intermediate results generated by Map tasks on the same node as the data blocks they process to reduce network transmission;

[0071] Task Scheduling: The MapReduce framework is responsible for scheduling tasks, ensuring that all tasks are executed in parallel on multiple computing nodes to improve the overall processing speed.

[0072] 2. Combining batch processing with real-time processing: Lambda architecture

[0073] Lambda architecture: including

[0074] Batch processing layer: performs batch processing calculations, usually periodically, and the calculation results are stored in the data warehouse;

[0075] Real-time processing layer: processes newly arrived data in real time and provides near real-time query results;

[0076] Service layer: merges the results of the batch processing layer and the real-time processing layer to provide the final query view.

[0077] The calculation method of this architecture is:

[0078] Batch processing layer: using databases or data warehouses to store and process large amounts of historical data;

[0079] Real-time processing layer: Use stream processing frameworks such as Apache Kafka Streams or Apache Flink to process real-time data streams;

[0080] Service layer: combines the results of the batch processing layer and the real-time processing layer to provide a unified query view.

[0081] 3. Stream processing: Apache Kafka Streams or Apache Flink

[0082] 3.1、Apache Kafka Streams:

[0083] Data stream processing: Process real-time data streams on the Kafka cluster;

[0084] Stream computing: Real-time processing and analysis of data streams.

[0085] 3.2、Apache Flink:

[0086] Stream processing framework: used to process and analyze real-time data streams;

[0087] State management: Provides a state management mechanism to support state persistence and fault tolerance.

[0088] The calculation method is:

[0089] Data stream processing: real-time processing of data streams, supporting operations on data streams, such as filtering, conversion, aggregation, etc.;

[0090] State management: manages state during stream processing, supports state persistence and fault tolerance;

[0091] Event time processing: Supports event time processing and can process time series data in data streams.

[0092] Through the models and calculation methods of these processing modes, the data processing module can efficiently process multiple data sets and realize real-time monitoring and early warning of the cardiovascular monitoring system.

[0093] In another optional embodiment, the difference between the patient private database and the public pathology database in the data storage module is that the data sets in the patient private database are matched one-to-one with the corresponding patient identity information, and an independent encrypted micro private database is matched one-to-one within the patient private database according to the patient identity information.

[0094] In this embodiment, the design of the patient private database is intended to ensure the security and confidentiality of the patient's private data. Specifically, each data set in the patient private database is closely associated with its corresponding patient identity information. This one-to-one matching mechanism can ensure the accuracy and reliability of the data. In addition, the data sets in the patient private database are also independently encrypted and stored in the micro private database. This encryption mechanism can effectively prevent unauthorized access and data leakage, thereby protecting the privacy and security of patients.

[0095] In another optional embodiment, the fusion model is a deep belief network, which is formed by stacking multiple restricted Boltzmann machines (RBMs).

[0096] In this embodiment, a deep belief network is used in a multimodal data fusion unit to fuse preprocessed data sets of different modalities to generate a comprehensive feature vector. A deep belief network (DBN) is a deep learning model that learns the feature representation of data by stacking multiple restricted Boltzmann machines (RBMs). Each RBM contains two layers: a visible layer and a hidden layer, which are connected by a weight matrix. During the training process, the RBM is trained by a contrastive divergence (CD) algorithm or a layer-by-layer pre-training method to learn the feature representation of the input data.

[0097] In this embodiment, the calculation method of the deep belief network is as follows:

[0098] 1. Input data preprocessing: First, preprocess the data sets from different modalities, including denoising, normalization, feature extraction and other steps to ensure the consistency and comparability of the data;

[0099] 2.RBM stacking: The preprocessed data set is input into the first RBM, which is trained by contrastive divergence algorithm or layer-by-layer pre-training method to learn the feature representation of the data. After the training is completed, the output of the first RBM is used as the input of the second RBM, and so on, until the last RBM;

[0100] 3. Feature representation learning: Through RBM stacking, the deep belief network learns the deep feature representation of the data. These feature representations can capture the complex patterns and relationships of the data, which helps to improve the accuracy of cardiovascular disease monitoring;

[0101] 4. Comprehensive feature vector generation: The output layer of the deep belief network takes the last layer output of the RBM stack as a comprehensive feature vector, which contains key information useful for cardiovascular disease monitoring;

[0102] Through the multimodal data fusion unit of the deep belief network, the cardiovascular monitoring system in this embodiment can effectively fuse data from different modalities to generate a comprehensive feature vector. This feature representation can better capture the complex characteristics of cardiovascular diseases, thereby improving the accuracy and reliability of monitoring.

[0103] In another optional embodiment, the display and warning module includes a threshold setting submodule, which sets individualized physiological parameter thresholds in combination with statistical analysis methods, and generates parameter threshold standards through a threshold adjustment mechanism.

[0104] In this embodiment, in another optional embodiment, the display and warning module includes a threshold setting submodule, which sets individualized physiological parameter thresholds in combination with statistical analysis methods, and generates parameter threshold standards through a threshold adjustment mechanism.

[0105] In this embodiment, the function of the threshold setting submodule is to determine the individualized physiological parameter threshold in the display and warning module to achieve accurate warning of cardiovascular disease monitoring. The specific steps are as follows:

[0106] 1. Data collection: Collect patients' rehabilitation-related data from multiple terminal devices, including electrocardiogram, cardiac magnetic resonance imaging, cardiac auscultation, photoplethysmography, blood pressure monitoring, and blood biochemical indicators;

[0107] 2. Data preprocessing: classify and preprocess the collected data to ensure the consistency and comparability of the data;

[0108] 3. Statistical analysis: Use statistical analysis methods, such as minimum-maximum standardization, median standardization, etc., to perform statistical analysis on the preprocessed data;

[0109] 4. Threshold setting: Combined with the statistical analysis results, individualized physiological parameter thresholds are set. These thresholds are dynamically adjusted according to the patient's specific health status and monitoring data to adapt to individual differences of patients;

[0110] 5. Threshold adjustment mechanism: Through the threshold adjustment mechanism, the threshold is dynamically adjusted according to the real-time changes of the monitoring data. This adjustment mechanism can ensure the accuracy and timeliness of the threshold, thereby improving the accuracy of the early warning;

[0111] 6. Generation of parameter threshold standards: Through the threshold setting submodule, parameter threshold standards suitable for cardiovascular monitoring are generated. These standards can be used to guide subsequent monitoring and early warning work.

[0112] Through the above steps, the display and warning module in this embodiment can set individualized physiological parameter thresholds according to the patient's specific health status and monitoring data, and generate parameter threshold standards through the threshold adjustment mechanism. This method helps to improve the accuracy and reliability of cardiovascular disease monitoring and provide patients with timely and accurate warning information.

[0113] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0114] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A cardiovascular monitoring system based on artificial intelligence, comprising a terminal device and a central processing system for data transmission with the terminal device, characterized in that: The central processing system includes Data collection module: collects patients' rehabilitation-related data from various terminal devices and generates data sets. The collected data include but are not limited to electrocardiogram collection, cardiac magnetic resonance imaging collection, cardiac auscultation collection, photoplethysmography collection, blood pressure monitoring and blood biochemical index collection; Data processing module: classifies the data sets obtained by different collection methods and selects specific processing strategies according to different data set types; Data storage module: partitions and stores the data sets processed by the data processing module, and collects the data into the public pathology database and the patient private database; Multimodal data fusion unit: uses a fusion model to fuse preprocessed data sets of different modalities to generate a comprehensive feature vector; Artificial intelligence analysis unit: uses pre-trained adaptive learning algorithms to analyze comprehensive feature vectors to achieve cardiovascular disease monitoring and early warning; Privacy protection unit: encrypts and de-identifies data to ensure data security; Display and warning module: displays the analysis results to the user in a graphical interface and issues warnings when abnormal conditions are detected; Emergency processing module: classifies warning signals and provides emergency processing plans when the warning signals exceed the set threshold.

2. The artificial intelligence-based cardiovascular monitoring system according to claim 1, characterized in that: The terminal device is a fixed monitoring device used by medical institutions or a wearable device that patients can carry with them for convenient use. The terminal device is provided with a data acquisition unit, a data output and receiving unit, an early warning prompt unit and a user information management unit.

3. The artificial intelligence-based cardiovascular monitoring system according to claim 2, characterized in that: The user information management unit stores the user's personal identity information, the user's historical health data set, and a future model of the disease condition generated by a medical model and continuously adjusted based on the user's health data. The user management unit can adjust the user's identity information, emergency contacts, health data, and warning thresholds.

4. The artificial intelligence-based cardiovascular monitoring system according to claim 2, characterized in that: When the terminal device is a wearable device, the terminal device is installed with a lightweight algorithm data processing system; at the same time, the terminal device also includes a remote medical service module, which automatically matches the medical institution with corresponding processing authority based on the warning level and emergency treatment plan and conducts active calls and medical consultations.

5. The artificial intelligence-based cardiovascular monitoring system according to claim 1, characterized in that: The data processing module classifies the data set according to the data type and source, and uses denoising, normalization and feature extraction methods when processing time series data; Image enhancement, segmentation, and feature extraction are used when processing image data; Data cleaning and data transformation methods are used when processing structured data.

6. The artificial intelligence-based cardiovascular monitoring system according to claim 5, characterized in that: The data processing module adopts parallel processing or stream processing or a combination of batch processing and real-time processing when processing multiple data sets.

7. The artificial intelligence-based cardiovascular monitoring system according to claim 6, characterized in that: When parallel processing is used, the MapReduce model is used; when batch processing is combined with real-time processing, the Lambda architecture is used; when stream processing is used, Apache Kafka Streams or Apache Flink is used.

8. The artificial intelligence-based cardiovascular monitoring system according to claim 1, characterized in that: The difference between the patient private database and the public pathology database in the data storage module is that the data sets in the patient private database are all matched one by one with the corresponding patient identity information, and the patient private database is matched one by one with an independent encrypted micro private database according to the patient identity information.

9. The artificial intelligence-based cardiovascular monitoring system according to claim 1, characterized in that: The fusion model is a deep belief network, which is formed by stacking multiple restricted Boltzmann machines (RBMs).

10. The artificial intelligence-based cardiovascular monitoring system according to claim 1, characterized in that: The display and warning module includes a threshold setting submodule, which sets individualized physiological parameter thresholds in combination with a statistical analysis method, and generates parameter threshold standards through a threshold adjustment mechanism.

Citation Information

Cited By

  • Cardiovascular disease full-cycle diagnosis and treatment management system based on multi-modal data fusion

    CN121687423A

  • Cleaning unit for cleaning a housing of a medical imaging device that surrounds a patient-accommodating region

    US20250170624A1