Data monitoring system for chronic cardiovascular disease

Abstract

The invention relates to the technical field of medical data monitoring, in particular to a data monitoring system for chronic cardiovascular diseases. The system comprises a patient physiological index data flow acquisition module, a physiological index change feature vector conversion module, a physiological index feature collapse anomaly quantification module and a patient physiological index anomaly response early warning module. Different physiological index data can be collected in real time through intelligent wearable equipment, stream data fusion processing and baseline correction processing are carried out, time sequence synchronization processing and physiological feature vector conversion processing are further carried out, and real-time physiological index change feature vectors of a patient are obtained; and acquiring the physiological index data of the patient in the historical chronic cardiovascular case, performing feature collapse anomaly quantitative calculation and anomaly response processing, generating a physiological index change anomaly response signal of the patient, and acting on the intelligent wearable device to execute the corresponding patient physiological index anomaly early warning prompt work. According to the invention, comprehensive and real-time monitoring management of chronic cardiovascular disease patients can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data monitoring, and in particular to a data monitoring system for chronic cardiovascular disease. Background Art

[0002] In the management of chronic cardiovascular disease, real-time data monitoring is crucial for early detection and long-term management of the disease. Through the integration of multiple sensors, including ECG sensors, photoelectric volume pulse wave sensors (PPG), blood pressure monitoring sensors, etc., smart wearable devices such as smart watches or health bracelets can simultaneously obtain heart rate, blood pressure, electrocardiogram and other physiological indicators. The integration of these sensors improves the comprehensiveness and accuracy of data collection, and the built-in data processing unit can process the data collected by the sensors in real time, apply advanced algorithms to analyze physiological signals, detect abnormal ECG changes, blood pressure fluctuations and other potential problems, and provide instant feedback to users with health warnings. However, traditional cardiovascular data monitoring methods mainly include regular hospital examinations and manual records. Although this method can provide certain health data, it has the problems of incomplete monitoring data and inaccurate analysis, which leads to inefficient management of patients with chronic cardiovascular disease. Summary of the invention

[0003] Based on this, it is necessary for the present invention to provide a data monitoring system for chronic cardiovascular disease to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a data monitoring system for chronic cardiovascular disease includes the following modules:

[0005] The patient physiological index data stream acquisition module is used to collect the real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease through smart wearable devices in real time, and perform stream data fusion processing on the real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease to obtain the real-time data stream of the patient's physiological index; the real-time data stream of the patient's physiological index is sent and transmitted to the data processing unit through the wireless transmission module;

[0006] The physiological index change characteristic vector conversion module is used to perform physiological index baseline correction processing on the patient's physiological index real-time data stream through the data processing unit to obtain the patient's physiological index baseline correction data stream; perform time series synchronization processing on each physiological index in the patient's physiological index baseline correction data stream to obtain the patient's physiological index data stream under the same time series dimension; perform physiological characteristic vector conversion processing on the patient's physiological index data stream under the same time series dimension to obtain the patient's real-time physiological index change characteristic vector;

[0007] The physiological index characteristic collapse abnormality quantification module is used to obtain the physiological index data of patients with historical chronic cardiovascular cases, and perform consistent physiological characteristic vector conversion on the physiological index data of patients with historical chronic cardiovascular cases to obtain the physiological index change characteristic vector of the patients with historical cases; based on the physiological index change characteristic vector of the patients with historical cases, the physiological index characteristic collapse abnormality quantification calculation is performed on the patient's real-time physiological index change characteristic vector to obtain the patient's real-time physiological index change characteristic collapse abnormality value;

[0008] The patient physiological indicator abnormal response warning module is used to perform physiological indicator abnormal response processing on the corresponding patient physiological indicators in the patient's real-time physiological indicator change feature vector based on the abnormal value of the patient's real-time physiological indicator change characteristic collapse, so as to generate a patient physiological indicator change abnormal response signal; the patient physiological indicator change abnormal response signal is applied to the smart wearable device to execute the corresponding patient physiological indicator abnormal warning prompt work.

[0009] Furthermore, the patient physiological index data stream acquisition module includes the following functions:

[0010] Collect real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease through smart wearable devices;

[0011] The real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease are subjected to stream data mapping and conversion through the stream data processing architecture to generate a sub-stream of heart rate data of patients with chronic cardiovascular disease, a sub-stream of ECG signal data of patients with chronic cardiovascular disease, a sub-stream of blood oxygen saturation data of patients with chronic cardiovascular disease and a sub-stream of blood pressure data of patients with chronic cardiovascular disease;

[0012] Performing stream data fusion processing on the heart rate data substream, the ECG signal data substream, the blood oxygen saturation data substream and the blood pressure data substream of the chronic cardiovascular disease patients to obtain the real-time data stream of the patients' physiological indicators;

[0013] The real-time data stream of the patient's physiological indicators is sent and transmitted to the data processing unit through the wireless transmission module.

[0014] Furthermore, the stream data fusion processing of the heart rate data substream of the chronic cardiovascular disease patient, the ECG signal data substream of the chronic cardiovascular disease patient, the blood oxygen saturation data substream of the chronic cardiovascular disease patient, and the blood pressure data substream of the chronic cardiovascular disease patient includes:

[0015] Performing data stream hierarchical extraction processing on the heart rate data substream of chronic cardiovascular disease patients, the ECG signal data substream of chronic cardiovascular disease patients, the blood oxygen saturation data substream of chronic cardiovascular disease patients, and the blood pressure data substream of chronic cardiovascular disease patients to obtain hierarchical physiological indicator data streams of chronic cardiovascular disease patients;

[0016] Performing statistical analysis on the time series characteristics of the data stream of stratified physiological indicators of patients with chronic cardiovascular disease to obtain a data set of time series characteristics of stratified physiological indicators of patients with chronic cardiovascular disease;

[0017] Performing feature smoothing and enhancement processing on the time series features of each physiological indicator data substream in the time series feature data set of stratified physiological indicators of patients with chronic cardiovascular disease, so as to obtain the enhanced time series feature data set of stratified physiological indicators of patients;

[0018] The temporal correlation mining analysis is performed on the temporal correlation between different physiological indicator temporal characteristics in the patient-stratified physiological indicator enhanced temporal characteristic data set to obtain the temporal correlation relationship between different physiological indicator temporal characteristics of chronic cardiovascular disease patients; based on the temporal correlation relationship between different physiological indicator temporal characteristics of chronic cardiovascular disease patients, the correlation feature map is connected and constructed between different physiological indicator temporal characteristics in the patient-stratified physiological indicator enhanced temporal characteristic data set to generate a dynamic correlation feature map of patient physiological indicators;

[0019] Based on the dynamic correlation feature map of patients' physiological indicators, the heart rate data substream of chronic cardiovascular disease patients, the ECG signal data substream of chronic cardiovascular disease patients, the blood oxygen saturation data substream of chronic cardiovascular disease patients and the blood pressure data substream of chronic cardiovascular disease patients are fused to obtain the real-time data stream of patients' physiological indicators.

[0020] Furthermore, the physiological index change feature vector conversion module includes the following functions:

[0021] Performing low-frequency interference elimination processing on the real-time data stream of the patient's physiological indicators through a high-pass filter in the data processing unit to obtain a low-frequency interference eliminated data stream of the patient's physiological indicators;

[0022] Performing high-frequency noise elimination processing on the low-frequency interference elimination data stream of the patient's physiological indicators through a low-pass filter in the data processing unit to obtain a high-frequency noise elimination data stream of the patient's physiological indicators;

[0023] Performing physiological index baseline correction processing on the high-frequency denoised data stream of the patient's physiological index to obtain a physiological index baseline correction data stream of the patient;

[0024] Performing time series synchronization processing on each physiological indicator in the patient's physiological indicator baseline correction data stream to obtain the patient's physiological indicator data stream in the same time series dimension;

[0025] The physiological indicator data stream of the patient in the same time series dimension is converted into a physiological feature vector to obtain a feature vector of the patient's real-time physiological indicator change.

[0026] Furthermore, the physiological index baseline correction processing of the high-frequency noise-removing data stream of the patient's physiological index includes:

[0027] Performing frequency domain analysis on the high-frequency denoised data stream of the patient's physiological indicators to generate a frequency distribution spectrum of the patient's physiological indicator signals;

[0028] Based on the frequency distribution spectrum of the patient's physiological indicator signal, a physiological indicator baseline drift detection is performed on the corresponding patient's physiological indicator high-frequency noise-removing data stream to obtain the patient's physiological indicator signal distribution baseline drift trend data;

[0029] Performing baseline drift correction analysis on the baseline drift trend data of the patient's physiological indicator signal distribution to obtain the baseline drift correction index of the patient's physiological indicator signal distribution;

[0030] Based on the baseline drift correction index of the patient's physiological indicator signal distribution, the corresponding patient's physiological indicator high-frequency denoised data stream is subjected to data stream baseline correction and adjustment processing to obtain the patient's physiological indicator baseline correction data stream.

[0031] Furthermore, the physiological characteristic vector conversion process of the patient's physiological indicator data stream in the same time series dimension includes:

[0032] Performing statistical analysis on the variability characteristics of the heart rate of patients with chronic cardiovascular disease in the patient physiological indicator data stream in the same time series dimension to obtain the heart rate variability characteristics of patients with chronic cardiovascular disease;

[0033] Performing signal waveform feature statistical analysis on ECG signals of patients with chronic cardiovascular disease in the patient physiological indicator data stream under the same time series dimension to obtain ECG signal waveform features of patients with chronic cardiovascular disease;

[0034] Perform blood oxygen activity level feature analysis on the blood oxygen saturation of chronic cardiovascular disease patients in the patient physiological indicator data stream in the same time series dimension to obtain the blood oxygen activity level feature of chronic cardiovascular disease patients;

[0035] The abnormal fluctuation characteristics of blood pressure of chronic cardiovascular disease patients in the patient physiological indicator data stream under the same time series dimension are analyzed to obtain the abnormal fluctuation characteristics of blood pressure of chronic cardiovascular disease patients;

[0036] The heart rate variability characteristics of chronic cardiovascular disease patients, the ECG signal waveform characteristics of chronic cardiovascular disease patients, the blood oxygen activity level characteristics of chronic cardiovascular disease patients, and the abnormal blood pressure fluctuation characteristics of chronic cardiovascular disease patients are converted into physiological feature vectors to obtain the feature vectors of real-time physiological index changes of patients.

[0037] Furthermore, the physiological indicator characteristic collapse abnormality quantification module includes the following functions:

[0038] Obtain physiological index data of patients with historical chronic cardiovascular cases;

[0039] Perform consistent physiological characteristic vector conversion on the physiological indicator data of historical chronic cardiovascular patients to obtain the characteristic vector of physiological indicator changes of historical patients;

[0040] Performing feature vector space synchronization mapping processing on the feature vectors of changes in physiological indicators of historical cases and the feature vectors of changes in real-time physiological indicators of patients, so as to obtain the feature vectors of physiological indicators of historical medical records and the feature vectors of real-time physiological indicators in the same feature vector space dimension;

[0041] Based on the physiological indicator feature vectors of historical medical records in the same feature vector space dimension, the corresponding real-time physiological indicator feature vectors are quantitatively calculated for the abnormal collapse of physiological indicator features to obtain the abnormal collapse value of the patient's real-time physiological indicator change features.

[0042] Furthermore, the physiological indicator characteristic collapse abnormality quantitative calculation of the corresponding real-time physiological indicator characteristic vector based on the historical medical record physiological indicator characteristic vector in the same characteristic vector space dimension includes:

[0043] Performing feature space distance measurement calculation between each corresponding physiological indicator feature in the historical medical record physiological indicator feature vector and the real-time physiological indicator feature vector under the same feature vector space dimension, so as to obtain a feature space distance measurement value between each historical physiological indicator feature vector and the corresponding real-time physiological indicator feature vector;

[0044] Performing feature vector space benchmark mapping analysis on the feature vectors of historical medical records physiological indicators and the feature vectors of real-time physiological indicators in the same feature vector space dimension to obtain a physiological indicator feature vector space benchmark mapping matrix;

[0045] Based on the physiological index characteristic vector space reference mapping matrix, the characteristic space reconstruction dimension difference analysis is performed on the real-time physiological index characteristic vectors in the same characteristic vector space dimension to obtain the patient's real-time physiological index characteristic vector space reconstruction dimension difference value;

[0046] Based on the feature space distance measurement value between each historical physiological indicator feature vector and the corresponding real-time physiological indicator feature vector and the difference value of the spatial reconstruction dimension of the patient's real-time physiological indicator feature vector, the physiological indicator feature collapse abnormality calculation formula is used to quantitatively calculate the physiological indicator feature collapse abnormality of the corresponding real-time physiological indicator feature vector to obtain the patient's real-time physiological indicator change feature collapse abnormality value.

[0047] Furthermore, the calculation formula for abnormal collapse of physiological index characteristics is specifically:

[0048]

[0049] ΔD(t)=U T *(X r,i (t)―X l,i (t));

[0050] In the formula, E c is the abnormal value of the collapse of the characteristic of the patient's real-time physiological index change, t1 is the initial time variable of integration, t2 is the end time variable of integration, t is the integral time variable parameter, n is the total number of physiological index feature vectors, i is the item measurement parameter of the physiological index feature vector, X s,i (t) is the i-th historical physiological index feature vector at time t in the same feature vector space dimension, X l,i (t) is the i-th real-time physiological index feature vector corresponding to time t in the same feature vector space dimension, ‖·‖ 2 is the feature space distance measurement paradigm, σ i is the standard deviation of the distance between the ith physiological index feature vectors, α is the feature vector space distance weighting coefficient, ΔD(t) is the difference value of the spatial reconstruction dimension of the patient's real-time physiological index feature vector at time t, β is the spatial reconstruction dimension difference weighting coefficient, U is the physiological index feature vector space reference mapping matrix, T is the matrix transposition symbol, X r,o (t) is the i-th real-time physiological index feature vector at time t in the original spatial dimension, and η is the correction coefficient of the abnormal value of the collapse of the patient's real-time physiological index change feature.

[0051] Furthermore, the patient physiological index abnormal response warning module includes the following functions:

[0052] Compare and judge the abnormal collapse value of the patient's real-time physiological indicator change characteristic according to the preset abnormal collapse threshold of the physiological indicator characteristic. When the abnormal collapse value of the patient's real-time physiological indicator change characteristic is greater than or equal to the preset abnormal collapse threshold of the physiological indicator characteristic, perform abnormal feature mapping processing on the corresponding patient's real-time physiological indicator change characteristic vector to generate an abnormal feature mapping map of the patient's physiological indicator; when the abnormal collapse value of the patient's real-time physiological indicator change characteristic is less than the preset abnormal collapse threshold of the physiological indicator characteristic, no processing is performed;

[0053] Based on the abnormal characteristic mapping graph of the patient's physiological indicators, abnormal physiological indicator response processing is performed on the corresponding patient's physiological indicators in the patient's real-time physiological indicator change characteristic vector to generate an abnormal physiological indicator change response signal of the patient;

[0054] The abnormal response signal of the patient's physiological index change is sent through the wireless transmission module and acts on the early warning controller in the smart wearable device to perform abnormal early warning control processing, generate the patient's physiological index abnormal early warning prompt control instruction, so as to execute the corresponding patient's physiological index abnormal early warning prompt work.

[0055] Beneficial effects of the present invention:

[0056] The data monitoring system for chronic cardiovascular disease proposed by the present invention, compared with the prior art, is composed of a patient physiological index data stream acquisition module, a physiological index change feature vector conversion module, a physiological index feature collapse abnormality quantification module and a patient physiological index abnormal response warning module as a whole. Its beneficial effect is that different physiological index data such as heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease are collected in real time by using smart wearable devices. The key to this step is that the physiological state of patients with chronic cardiovascular disease can be continuously and accurately monitored. Non-invasive data collection can be performed in the daily life of patients without additional medical intervention, thereby improving the comfort of patients and the frequency of data collection. This real-time monitoring is particularly important for patients with chronic cardiovascular disease, because their health status will change in a short period of time. Through continuous monitoring, abnormal data can be discovered in time, and potential health problems can be warned in advance, which helps doctors and patients to respond quickly and take necessary medical measures, thereby reducing the risk of acute onset of the disease, thereby improving the comprehensiveness of physiological data monitoring of patients with chronic cardiovascular disease. In addition, by performing streaming data fusion processing on the real-time physiological indicator data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease, the key of this step is to be able to be reflected in the comprehensive analysis of multiple data dimensions and provide a comprehensive health monitoring view. Data fusion can combine individual physiological indicators to discover potential health trends and associations. For example, the combination of heart rate and blood pressure data can reveal the overall condition of vascular health, while the fusion of ECG signal and heart rate can help analyze potential arrhythmia problems. Comprehensive data analysis can not only provide more accurate health assessments, but also improve the effectiveness of early warnings through pattern recognition and anomaly detection, helping doctors better understand the health status of patients and develop personalized abnormal warning work plans. At the same time, the real-time data stream of the patient's physiological indicators is sent to the data processing unit through the use of the built-in wireless transmission module. The key to this step is to realize the remote real-time transmission of the integrated data stream, thereby improving the flexibility and response speed of data processing. The wireless transmission module enables the patient data to be quickly and seamlessly transmitted to the data processing unit, avoiding the inconvenience caused by traditional wired connections. Real-time data transmission enables doctors and medical staff to obtain the patient's health information instantly and conduct remote monitoring. This timely data transmission not only improves the quality of patient care, thereby saving time and cost, but also supports rapid data upload in emergency situations, thereby enhancing the ability to respond to sudden health events in patients with chronic cardiovascular disease.Secondly, the real-time data stream of the patient's physiological indicators is processed by using a data processing unit to perform physiological indicator baseline correction. The beneficial effect of this step is to correct the data baseline drift caused by time, equipment deviation or measurement condition changes. The physiological indicator data will have baseline offset at different time points or under different conditions, thereby affecting the accuracy of the data. The baseline correction process can standardize the data so that each data point is compared with a unified reference point, thereby eliminating the deviation caused by equipment or environmental changes. This process can ensure the long-term consistency of the data, improve the comparability of the data, provide doctors with more reliable health monitoring information, and help make more scientific decisions. In addition, each physiological indicator in the patient's physiological indicator baseline correction data stream is processed in time series synchronization. The key to this step is to enable different physiological indicators to be compared and analyzed on the same time axis. Different physiological indicators (such as heart rate, blood oxygen saturation, etc.) have different sampling frequencies and timestamps. The time series synchronization process aligns these indicators to ensure that all data are reflected in the same time dimension, helping doctors better understand the relationship between various physiological parameters, thereby improving the overall grasp of the patient's health status. By converting the physiological indicator data stream of patients in the same time series dimension into physiological feature vectors, multi-dimensional physiological data can be converted into analyzable feature vectors and key physiological change information can be extracted. This conversion can compress complex time series data into feature vectors that are easier to process and interpret, which is convenient for machine learning algorithms or other analysis methods to perform further analysis. Through the generation of feature vectors, the changing trends and abnormal patterns of physiological indicators can be better captured, thereby improving the accuracy of physiological data analysis and abnormal management efficiency of patients with chronic cardiovascular disease. This not only optimizes the data analysis process, but also provides strong data support for the realization of precision medicine and personalized health management. Then, by obtaining the physiological indicator data of historical chronic cardiovascular cases, the key to this step is to provide a valuable reference benchmark for comparing and analyzing the health status of current patients. Historical data not only includes the long-term health trends of patients, but also covers data and changes in the disease at different stages. By collecting these historical data, the development process, reaction and health trend of the patient's disease can be better understood. The acquisition of historical case data lays a solid foundation for subsequent data analysis, so that more subtle changes and abnormalities can be found when comparing real-time data. The physiological indicator data of historical chronic cardiovascular patients were also converted into consistent physiological feature vectors. The key to this processing process is to standardize the historical data into a feature representation consistent with the previous real-time monitoring data, so that the data can be effectively compared and analyzed. The feature vector conversion can simplify complex physiological data into a vector form that is easier to handle, retaining the key features of the data while eliminating differences in data under different measurement conditions and time points.Through consistent feature vector conversion, it can be ensured that the comparison basis of historical case data is unified during analysis, thereby improving the reliability and consistency of the subsequent analysis process. In addition, by performing quantitative calculation of the collapse abnormality of physiological indicators based on the feature vector of physiological indicator changes of patients in historical cases, the key to this processing step is to quantify the degree of abnormality of real-time data relative to historical data and provide specific abnormal value indicators for clinical judgment. The quantitative calculation of physiological indicator feature collapse abnormality can reveal significant changes or abnormal trends in the patient's current health status by comparing real-time data with the standards of historical data. This quantitative calculation helps to identify potential health problems that have been neglected, allowing doctors to take timely intervention measures to prevent further deterioration or acute attacks of the disease. By providing clear abnormal values, doctors can more accurately adjust health decision-making plans and health management strategies, improve the pertinence and effectiveness of medical intervention, and thus improve the health prognosis of patients with chronic cardiovascular disease. Finally, the abnormal physiological indicator response processing is performed on the corresponding patient physiological indicators in the feature vector of the patient's real-time physiological indicator change based on the abnormal value of the patient's real-time physiological indicator change characteristic obtained by previous quantitative calculation. This step significantly improves the response ability to abnormal situations and can more accurately identify which physiological indicators have undergone abnormal changes. This processing method makes the abnormal response processing more accurate and targeted. The generated abnormal response signal can not only reflect the changes in health status in a timely manner, but also provide medical personnel with detailed information for further analysis and intervention. This step helps to locate problems more quickly, reduce the health risks faced by patients, and provide a basis for formulating personalized abnormal risk management plans. In addition, the abnormal response signal of the patient's physiological indicator changes is sent to the smart wearable device through the use of a wireless transmission module, thereby generating an abnormal warning prompt control instruction to execute the corresponding abnormal warning work of the patient's physiological indicators. This process greatly improves the effectiveness and response speed of real-time warning. The use of the wireless transmission module ensures the instant transmission of data without the need for physical connection or human intervention, reducing data transmission delays. This enables the warning system to respond quickly when an abnormal situation occurs and send a corresponding warning signal to the wearable device. After receiving the signal, the warning controller in the smart wearable device will generate a warning prompt control instruction to promptly remind patients and medical staff of potential health risks. Such an automated warning mechanism not only improves patients' timely understanding of their own health status, but also enables the medical team to take action as soon as possible to prevent potential health problems. This real-time warning and feedback mechanism greatly improves the efficiency and accuracy of health management, and helps to improve patients' prognosis and quality of life. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings:

[0058] Figure 1 is a schematic diagram of a data monitoring system for chronic cardiovascular disease according to the present invention;

[0059] Figure 2 for Figure 1 Schematic diagram of the functional flow of the patient physiological index data stream acquisition module;

[0060] Figure 3 for Figure 2 Schematic diagram of the functional flow of the physiological index change feature vector conversion module. DETAILED DESCRIPTION

[0061] The technical system of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0062] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor systems and / or microcontroller systems.

[0063] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0064] To achieve this, please refer to Figures 1 to 3 The present invention provides a data monitoring system for chronic cardiovascular disease, the system comprising the following modules:

[0065] The patient physiological index data stream acquisition module is used to collect the real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease through smart wearable devices in real time, and perform stream data fusion processing on the real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease to obtain the real-time data stream of the patient's physiological index; the real-time data stream of the patient's physiological index is sent and transmitted to the data processing unit through the wireless transmission module;

[0066] The physiological index change characteristic vector conversion module is used to perform physiological index baseline correction processing on the patient's physiological index real-time data stream through the data processing unit to obtain the patient's physiological index baseline correction data stream; perform time series synchronization processing on each physiological index in the patient's physiological index baseline correction data stream to obtain the patient's physiological index data stream under the same time series dimension; perform physiological characteristic vector conversion processing on the patient's physiological index data stream under the same time series dimension to obtain the patient's real-time physiological index change characteristic vector;

[0067] The physiological index characteristic collapse abnormality quantification module is used to obtain the physiological index data of patients with historical chronic cardiovascular cases, and perform consistent physiological characteristic vector conversion on the physiological index data of patients with historical chronic cardiovascular cases to obtain the physiological index change characteristic vector of the patients with historical cases; based on the physiological index change characteristic vector of the patients with historical cases, the physiological index characteristic collapse abnormality quantification calculation is performed on the patient's real-time physiological index change characteristic vector to obtain the patient's real-time physiological index change characteristic collapse abnormality value;

[0068] The patient physiological indicator abnormal response warning module is used to perform physiological indicator abnormal response processing on the corresponding patient physiological indicators in the patient's real-time physiological indicator change feature vector based on the abnormal value of the patient's real-time physiological indicator change characteristic collapse, so as to generate a patient physiological indicator change abnormal response signal; the patient physiological indicator change abnormal response signal is applied to the smart wearable device to execute the corresponding patient physiological indicator abnormal warning prompt work.

[0069] In the embodiment of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of a data monitoring system for chronic cardiovascular disease according to the present invention. In this example, the data monitoring system for chronic cardiovascular disease includes the following modules:

[0070] S1: Patient physiological index data stream acquisition module, used for collecting real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of chronic cardiovascular disease patients through smart wearable devices, and performing stream data fusion processing on the real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of chronic cardiovascular disease patients to obtain the real-time data stream of patient physiological index; sending and transmitting the real-time data stream of patient physiological index to the data processing unit through the wireless transmission module;

[0071] In an embodiment of the present invention, by using smart wearable devices, such as a heart rate monitor, an ECG sensor, a blood oxygen saturation detector, and a sphygmomanometer, these devices are embedded in the wearable device and can continuously monitor and record the patient's heart rate, ECG signal, blood oxygen saturation, and blood pressure. The built-in sensors of the device convert the collected physiological data into electrical signals, which are processed through an analog-to-digital converter (ADC) to convert the analog signals into digital data, thereby collecting and obtaining real-time physiological indicator data of heart rate, ECG signal, blood oxygen saturation, and blood pressure of patients with chronic cardiovascular disease. At the same time, a stream data processing architecture is used to map and convert the data streams of the real-time physiological indicator data of heart rate, ECG signal, blood oxygen saturation and blood pressure of the chronic cardiovascular disease patients collected above. This process uses a stream processing framework such as Apache Kafka or Apache Flink to independently map each physiological indicator data (heart rate, ECG signal, blood oxygen saturation, blood pressure) into a data sub-stream, and convert the original data into a standardized data format by using the corresponding parsing and conversion logic. This conversion includes timestamp marking, data unit standardization and data format normalization to ensure that the data of each indicator can be processed independently and has consistency, and the data stream fusion algorithm is used to perform data stream fusion processing on the previously converted chronic cardiovascular disease patient physiological indicator data sub-streams to integrate the heart rate data sub-stream, ECG signal data sub-stream, blood oxygen saturation data sub-stream and blood pressure data sub-stream into a unified real-time data stream. This process usually involves data alignment, data fusion technology such as weighted average, Kalman filtering or data fusion model application to integrate the information of each data sub-stream to obtain the patient's physiological indicator real-time data stream. Then, the real-time data stream of the patient's physiological indicators obtained after the previous fusion processing is sent to the data processing unit by using the built-in wireless transmission module (such as Wi-Fi, Bluetooth or Zigbee module) of the smart wearable device. The wireless transmission module is responsible for sending the fused data stream through the wireless network to ensure that the data transmission process from the wearable device to the processing unit is stable and reliable for further analysis, storage or visualization processing.

[0072] S2: A physiological indicator change characteristic vector conversion module, which is used to perform physiological indicator baseline correction processing on the patient's physiological indicator real-time data stream through a data processing unit to obtain the patient's physiological indicator baseline correction data stream; perform time series synchronization processing on each physiological indicator in the patient's physiological indicator baseline correction data stream to obtain the patient's physiological indicator data stream in the same time series dimension; perform physiological characteristic vector conversion processing on the patient's physiological indicator data stream in the same time series dimension to obtain the patient's real-time physiological indicator change characteristic vector;

[0073] In an embodiment of the present invention, a low-frequency interference elimination process is performed on a real-time data stream of a patient's physiological index previously acquired in real time by using a high-pass filter in a data processing unit, so that the real-time data stream is input into a high-pass filter, the cutoff frequency of the filter is set at a predetermined value to filter out signal components below the frequency, and a high-frequency noise elimination process is performed on the real-time data stream of the patient's physiological index previously obtained after low-frequency interference elimination by using a low-pass filter, so that the low-frequency interference eliminated data stream is input into the low-pass filter, so as to smooth the signal and filter out high-frequency interference (such as electrical noise or high-frequency interference), and output a high-frequency noise-eliminating data stream, which contains pure physiological index information after high-frequency noise is eliminated, and the corresponding baseline drift component is extracted from the high-frequency noise-eliminating data stream of the patient's physiological index, usually by calculating the moving average of the data stream or using a regression method, and by quantifying and calculating the corresponding baseline drift correction index from the detected baseline drift component, and subtracting it from the original data stream to correct the baseline of the data stream. This correction process can use a baseline correction method to ensure that the physiological indicator data is analyzed at a preset baseline level, so as to obtain a patient's physiological indicator baseline-corrected data stream. Then, by performing time series synchronization processing on each physiological indicator in the baseline correction data stream of the patient's physiological indicators obtained after the baseline correction, timestamp data is extracted from different physiological indicator data streams, and the time series of each data stream is aligned, so as to fill the time intervals between the data by using interpolation methods (such as linear interpolation or spline interpolation), ensuring that all physiological indicators have valid data at the same time point, and applying corresponding feature analysis methods to the patient's physiological indicator data stream in the same time series dimension obtained after the time series synchronization, so as to analyze and extract key physiological features, such as mean, standard deviation, frequency domain features, etc., and by using feature vector generation methods (such as principal component analysis PCA), the extracted feature data is converted into a physiological feature vector, which summarizes the change information of the patient's real-time physiological indicators, and finally obtains the patient's real-time physiological indicator change feature vector.

[0074] S3: A physiological index characteristic collapse abnormality quantification module is used to obtain the physiological index data of patients with historical chronic cardiovascular cases, and perform a consistent physiological characteristic vector conversion on the physiological index data of patients with historical chronic cardiovascular cases to obtain the physiological index change characteristic vector of the patients with historical cases; based on the physiological index change characteristic vector of the patients with historical cases, the physiological index characteristic collapse abnormality quantification calculation is performed on the patient's real-time physiological index change characteristic vector to obtain the patient's real-time physiological index change characteristic collapse abnormality value;

[0075] In an embodiment of the present invention, the physiological index data of historical chronic cardiovascular patients are obtained through a medical database or recording system. These data include but are not limited to heart rate, ECG signal, blood oxygen saturation and blood pressure, etc. These data should be extracted from the electronic health record (EHR) system to ensure that the extracted data has time series characteristics and contains the patient's historical physiological measurement records, so as to obtain the physiological index data of historical chronic cardiovascular patients. At the same time, by using feature engineering technology, the acquired physiological index data of historical chronic cardiovascular patients and the previous physiological index real-time data stream are converted into consistent feature vectors to achieve the conversion of the physiological index data of historical chronic cardiovascular patients into corresponding physiological feature vectors (including historical feature data such as heart rate variability, ECG signal waveform, blood oxygen activity level and abnormal blood pressure fluctuations), and by applying principal component analysis (PCA) or other dimensionality reduction techniques, the high-dimensional physiological data is mapped to a low-dimensional feature space, so as to obtain the characteristic vector of physiological index changes of historical patients. Then, the physiological indicator feature vectors of historical cases and the real-time physiological indicator feature vectors are mapped to the same feature space by using feature vector space synchronous mapping to ensure that their dimensions are consistent with the feature space of the historical case data, and by calculating the feature space distance measurement value between each pair of corresponding physiological indicator features, the Euclidean distance, Manhattan distance and other measurement methods can be used. The specific operation is to calculate the difference between each feature vector element of the historical medical record and the element at the corresponding position in the real-time physiological indicator feature vector, square these differences, sum them and then take the square root to obtain the feature space distance value, and also perform a benchmark mapping analysis of the feature vector space on the physiological indicator feature vectors of the historical medical records and the real-time physiological indicator feature vectors to calculate the benchmark mapping matrix that maps them to the same feature space, and feature the real-time physiological indicator feature vectors by combining the feature vector space benchmark mapping matrix obtained by the previous analysis. Differential analysis of spatial reconstruction dimension is performed to transform the real-time physiological indicator feature vector into a new feature space through a reference mapping matrix, and then compare it with the reconstruction of the feature vector in the original space to calculate the difference between the reconstructed and mapped feature vector and the original feature vector to obtain the corresponding vector space reconstruction dimension differential statistics, and construct a suitable calculation formula by combining the feature space distance value and the vector space reconstruction dimension differential statistics to perform quantitative calculation of the collapse abnormality of the physiological indicator feature vector for the corresponding real-time physiological indicator feature vector. This formula can be in the form of a weighted sum or a weighted difference to quantitatively calculate the collapse abnormality value of the corresponding physiological indicator feature vector. This collapse abnormality value reflects whether the change of the real-time physiological indicator deviates significantly from the normal range in the historical data, thereby helping to detect potential physiological abnormalities, and finally obtaining the collapse abnormality value of the patient's real-time physiological indicator change characteristics.

[0076] S4: A patient physiological indicator abnormal response warning module is used to perform physiological indicator abnormal response processing on the corresponding patient physiological indicators in the patient's real-time physiological indicator change feature vector based on the abnormal value of the patient's real-time physiological indicator change characteristic collapse, so as to generate a patient physiological indicator change abnormal response signal; the patient physiological indicator change abnormal response signal is applied to the smart wearable device to execute the corresponding patient physiological indicator abnormality warning prompt work.

[0077] In an embodiment of the present invention, by using a preset abnormal collapse threshold of physiological index characteristics, such as the normal fluctuation range of physiological data such as heart rate, respiratory rate, and body temperature, the abnormal collapse value of the patient's real-time physiological index change characteristics previously obtained by quantitative calculation is compared and judged. If the abnormal collapse value of the physiological index change characteristics collected in real time exceeds or is equal to the preset abnormal threshold, the abnormal feature mapping process is performed on the real-time physiological index feature vector corresponding to the patient. The data used in the mapping process may include specific abnormal identification algorithms, such as k-means clustering, support vector machine (SVM), etc., and the corresponding physiological index change feature vector is converted into an abnormal feature mapping spectrum, thereby generating a patient physiological index abnormal feature mapping spectrum. If the abnormal collapse value of the real-time feature does not reach the abnormal threshold, no processing is performed to save computing resources and avoid false alarms. At the same time, by combining the abnormal characteristic mapping map of the patient's physiological indicators drawn previously, the abnormal response processing of the physiological indicators of the patient corresponding to the characteristic vector of the patient's real-time physiological indicators change is performed. The analysis process of the map can use data mining technology, such as anomaly detection algorithm, to identify specific abnormal physiological indicators by marking the abnormal area in the map, and analyze the relationship between the abnormal characteristic map and the real-time physiological data by applying a response processing algorithm, such as dynamic time warping (DTW) or other matching algorithms. The generated abnormal response signal includes the real-time value of the corresponding abnormal indicator, the abnormal level and other information, thereby responding to the generation of an abnormal response signal of the patient's physiological indicator change. Then, the generated abnormal response signal of the patient's physiological index change is wirelessly transmitted by using the built-in wireless transmission module of the smart wearable device. The wireless transmission module can be Bluetooth, Wi-Fi or other wireless communication technologies to send the abnormal signal to the early warning controller in the smart wearable device. After receiving the signal, the early warning controller processes it through a preset control logic, such as setting threshold judgment, abnormal level classification, etc., so as to respond and generate corresponding control instructions, including visual, auditory or vibration reminder methods, etc., and send warnings to users or medical staff in real time through smart wearable devices. For example, the controller will activate the alarm sound on the smart wearable device or the red warning icon on the display screen, so as to deal with the abnormal situation in time and ensure that the patient can get quick intervention, thereby executing the corresponding abnormal early warning prompt of the patient's physiological index.

[0078] Furthermore, the patient physiological index data stream acquisition module includes the following functions:

[0079] Collect real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease through smart wearable devices;

[0080] The real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease are subjected to stream data mapping and conversion through the stream data processing architecture to generate a sub-stream of heart rate data of patients with chronic cardiovascular disease, a sub-stream of ECG signal data of patients with chronic cardiovascular disease, a sub-stream of blood oxygen saturation data of patients with chronic cardiovascular disease and a sub-stream of blood pressure data of patients with chronic cardiovascular disease;

[0081] Performing stream data fusion processing on the heart rate data substream, the ECG signal data substream, the blood oxygen saturation data substream and the blood pressure data substream of the chronic cardiovascular disease patients to obtain the real-time data stream of the patients' physiological indicators;

[0082] The real-time data stream of the patient's physiological indicators is sent and transmitted to the data processing unit through the wireless transmission module.

[0083] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 The functional flow diagram of the patient physiological index data stream acquisition module in the embodiment includes the following functions:

[0084] S11: Real-time collection of physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease through smart wearable devices;

[0085] In an embodiment of the present invention, by using smart wearable devices, such as a heart rate monitor, an ECG sensor, a blood oxygen saturation detector, and a sphygmomanometer, these devices are embedded in the wearable device and can continuously monitor and record the patient's heart rate, ECG signal, blood oxygen saturation, and blood pressure. The built-in sensors of the device convert the collected physiological data into electrical signals, which are processed through an analog-to-digital converter (ADC) to convert the analog signals into digital data, and finally the real-time physiological indicator data of the heart rate, ECG signal, blood oxygen saturation, and blood pressure of patients with chronic cardiovascular disease are collected.

[0086] S12: performing stream data mapping conversion on the real-time physiological index data of the heart rate, ECG signal, blood oxygen saturation and blood pressure of the chronic cardiovascular disease patient through the stream data processing architecture to generate a sub-stream of the heart rate data of the chronic cardiovascular disease patient, a sub-stream of the ECG signal data of the chronic cardiovascular disease patient, a sub-stream of the blood oxygen saturation data of the chronic cardiovascular disease patient and a sub-stream of the blood pressure data of the chronic cardiovascular disease patient;

[0087] In an embodiment of the present invention, a stream data processing architecture is used to map and convert the real-time physiological indicator data of heart rate, ECG signal, blood oxygen saturation and blood pressure of chronic cardiovascular disease patients collected above. This process uses a stream processing framework such as Apache Kafka or Apache Flink to independently map each physiological indicator data (heart rate, ECG signal, blood oxygen saturation, blood pressure) into a data sub-stream, and convert the original data into a standardized data format by using corresponding parsing and conversion logic. This conversion includes timestamp marking, data unit standardization and data format normalization to ensure that the data of each indicator can be processed independently and with consistency, and finally generate a heart rate data sub-stream of chronic cardiovascular disease patients, an ECG signal data sub-stream of chronic cardiovascular disease patients, a blood oxygen saturation data sub-stream of chronic cardiovascular disease patients and a blood pressure data sub-stream of chronic cardiovascular disease patients.

[0088] S13: performing stream data fusion processing on the heart rate data substream of the chronic cardiovascular disease patient, the ECG signal data substream of the chronic cardiovascular disease patient, the blood oxygen saturation data substream of the chronic cardiovascular disease patient, and the blood pressure data substream of the chronic cardiovascular disease patient to obtain a real-time data stream of the patient's physiological indicators;

[0089] In an embodiment of the present invention, a data stream fusion algorithm is used to perform data stream fusion processing on the previously converted heart rate data substream of a chronic cardiovascular disease patient, ECG signal data substream of a chronic cardiovascular disease patient, blood oxygen saturation data substream of a chronic cardiovascular disease patient, and blood pressure data substream of a chronic cardiovascular disease patient, so as to integrate the heart rate data substream, ECG signal data substream, blood oxygen saturation data substream and blood pressure data substream into a unified real-time data stream. This process usually involves data alignment and the application of data fusion techniques such as weighted averaging, Kalman filtering or a data fusion model to integrate the information of each data substream, and the fused data stream will provide a comprehensive real-time status of the patient's physiological indicators, and finally obtain a real-time data stream of the patient's physiological indicators.

[0090] S14: The real-time data stream of the patient's physiological indicators is sent and transmitted to the data processing unit via the wireless transmission module.

[0091] In an embodiment of the present invention, a real-time data stream of the patient's physiological indicators obtained after fusion processing is sent to a data processing unit by using a built-in wireless transmission module (such as a Wi-Fi, Bluetooth or Zigbee module) of a smart wearable device. The wireless transmission module is responsible for sending the fused data stream through a wireless network to ensure that the data transmission process from the wearable device to the processing unit is stable and reliable for further analysis, storage or visualization processing.

[0092] Furthermore, the stream data fusion processing of the heart rate data substream of the chronic cardiovascular disease patient, the ECG signal data substream of the chronic cardiovascular disease patient, the blood oxygen saturation data substream of the chronic cardiovascular disease patient, and the blood pressure data substream of the chronic cardiovascular disease patient includes:

[0093] Performing data stream hierarchical extraction processing on the heart rate data substream of chronic cardiovascular disease patients, the ECG signal data substream of chronic cardiovascular disease patients, the blood oxygen saturation data substream of chronic cardiovascular disease patients, and the blood pressure data substream of chronic cardiovascular disease patients to obtain hierarchical physiological indicator data streams of chronic cardiovascular disease patients;

[0094] In an embodiment of the present invention, the heart rate data substream, ECG signal data substream, blood oxygen saturation data substream and blood pressure data substream of a chronic cardiovascular disease patient are processed in layers by using a data stream management tool, so as to mark and classify these physiological indicator data according to timestamps and data types. The heart rate data substream will be processed separately in layers to extract heart rate variability, the ECG signal substream will be segmented by a signal processing tool to extract PQRST waveform features, and the blood oxygen saturation data substream and blood pressure data substream will be processed using data cleaning and normalization techniques to ensure the quality and consistency of each data stream, and finally obtain the layered physiological indicator data stream of chronic cardiovascular disease patients.

[0095] Preferably, a time series characteristic statistical analysis is performed on the data stream of stratified physiological indicators of patients with chronic cardiovascular disease to obtain a time series characteristic data set of stratified physiological indicators of patients with chronic cardiovascular disease;

[0096] In an embodiment of the present invention, a statistical analysis of the time series characteristics of the stratified physiological indicator data stream of chronic cardiovascular disease patients obtained after stratification is performed by using a time series analysis tool, so as to model each physiological indicator by applying a time series analysis algorithm such as an autoregressive integrated moving average (ARIMA) model, and by calculating the time series characteristics such as the mean, standard deviation, trend and seasonality of each time series data, finally a time series characteristic data set of stratified physiological indicators of chronic cardiovascular disease patients is obtained.

[0097] Preferably, the time series characteristics of each physiological indicator data substream in the chronic cardiovascular disease patient stratified physiological indicator time series characteristic data set are smoothed and enhanced to obtain the patient stratified physiological indicator enhanced time series characteristic data set;

[0098] In an embodiment of the present invention, the time series features of each physiological indicator data substream in the time series feature data set of stratified physiological indicators of chronic cardiovascular disease patients obtained by previous statistical analysis are smoothed and enhanced to reduce noise by using smoothing techniques such as sliding average or exponentially weighted smoothing (EWMA), and feature enhancement uses data enhancement technology, such as generative adversarial networks (GANs) or time series data interpolation methods, to expand the feature data set and improve the expressiveness of the features, ultimately obtaining a patient-stratified physiological indicator enhanced time series feature data set.

[0099] Preferably, a temporal correlation mining analysis is performed on different temporal characteristics of physiological indicators in the patient-stratified physiological indicator enhanced temporal characteristic data set to obtain the temporal correlation relationship between different temporal characteristics of physiological indicators of patients with chronic cardiovascular disease; based on the temporal correlation relationship between different temporal characteristics of physiological indicators of patients with chronic cardiovascular disease, a correlation feature map is constructed to connect different temporal characteristics of physiological indicators in the patient-stratified physiological indicator enhanced temporal characteristic data set to generate a dynamic correlation feature map of patient physiological indicators;

[0100] In an embodiment of the present invention, by using an association rule learning algorithm such as Apriori or FP-Growth, a time series association mining analysis is performed on different physiological indicator time series features in a patient-stratified physiological indicator enhanced time series feature data set obtained by a previous analysis, so as to discover the association pattern between different physiological indicators, and by calculating the correlation between each time series feature, a time series association rule relationship is generated to describe the mutual influence and dependency between physiological indicators, so as to obtain the time series association relationship between different physiological indicator time series features of patients with chronic cardiovascular disease. At the same time, by combining the time series association relationship between different physiological indicator time series features of patients with chronic cardiovascular disease obtained by previous mining analysis, a graph construction tool such as Gephi or Cytoscape is used to connect and construct an association graph between corresponding different physiological indicator time series features in the patient-stratified physiological indicator enhanced time series feature data set, so as to represent the time series features of different physiological indicators and their association relationships as a graph structure, in which nodes represent physiological indicator time series features, edges represent the association relationships between them, and finally a dynamic association feature graph of patient physiological indicators is generated to display the dynamic connection between each physiological indicator feature.

[0101] Preferably, based on the dynamic correlation characteristic map of the patients' physiological indicators, the heart rate data substream of the chronic cardiovascular disease patients, the ECG signal data substream of the chronic cardiovascular disease patients, the blood oxygen saturation data substream of the chronic cardiovascular disease patients and the blood pressure data substream of the chronic cardiovascular disease patients are subjected to stream data fusion processing to obtain the real-time data stream of the patients' physiological indicators.

[0102] In an embodiment of the present invention, the heart rate data substream of patients with chronic cardiovascular disease, the ECG signal data substream of patients with chronic cardiovascular disease, the blood oxygen saturation data substream of patients with chronic cardiovascular disease, and the blood pressure data substream of patients with chronic cardiovascular disease are fused by using the corresponding association relationship in the dynamic association feature map of the patient's physiological indicators previously constructed, so that the heart rate data substream, the ECG signal data substream, the blood oxygen saturation data substream and the blood pressure data substream are synthesized according to the corresponding data stream association relationship using a stream data fusion algorithm such as a weighted fusion or a multi-view learning method. During the fusion process, each data stream is synchronized through a timing alignment technology, and a weighted algorithm is applied to combine the timing characteristics of different indicators to generate a comprehensive real-time data stream, and finally a real-time data stream of the patient's physiological indicators is obtained.

[0103] Furthermore, the physiological index change feature vector conversion module includes the following functions:

[0104] Performing low-frequency interference elimination processing on the real-time data stream of the patient's physiological indicators through a high-pass filter in the data processing unit to obtain a low-frequency interference eliminated data stream of the patient's physiological indicators;

[0105] Performing high-frequency noise elimination processing on the low-frequency interference elimination data stream of the patient's physiological indicators through a low-pass filter in the data processing unit to obtain a high-frequency noise elimination data stream of the patient's physiological indicators;

[0106] Performing physiological index baseline correction processing on the high-frequency denoised data stream of the patient's physiological index to obtain a physiological index baseline correction data stream of the patient;

[0107] Performing time series synchronization processing on each physiological indicator in the patient's physiological indicator baseline correction data stream to obtain the patient's physiological indicator data stream in the same time series dimension;

[0108] The physiological indicator data stream of the patient in the same time series dimension is converted into a physiological feature vector to obtain a feature vector of the patient's real-time physiological indicator change.

[0109] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 FIG. 1 is a functional flow chart of a physiological indicator change feature vector conversion module in FIG. 1 . In this embodiment, the physiological indicator change feature vector conversion module includes the following functions:

[0110] S21: performing low-frequency interference elimination processing on the real-time data stream of the patient's physiological indicators through a high-pass filter in the data processing unit to obtain a low-frequency interference eliminated data stream of the patient's physiological indicators;

[0111] In an embodiment of the present invention, a low-frequency interference elimination process is performed on the real-time data stream of the patient's physiological indicators previously acquired in real time by using a high-pass filter in a data processing unit, so that the real-time data stream is input into the high-pass filter, and the cutoff frequency of the filter is set at a predetermined value to filter out signal components below the frequency, and by eliminating low-frequency interference (for example, baseline drift or slow fluctuation) therein, the high-frequency information in the physiological signal is retained. The processed data stream is a low-frequency interference eliminated data stream, which includes the physiological indicator data after the low-frequency interference is eliminated, and finally the patient's physiological indicator low-frequency interference eliminated data stream is obtained.

[0112] S22: performing high-frequency noise elimination processing on the low-frequency interference elimination data stream of the patient's physiological indicators through a low-pass filter in the data processing unit to obtain a high-frequency noise elimination data stream of the patient's physiological indicators;

[0113] In an embodiment of the present invention, a low-pass filter in a data processing unit is used to perform high-frequency denoising on a low-frequency interference-eliminated data stream of a patient's physiological indicators that has been previously obtained after low-frequency interference elimination, so that the low-frequency interference-eliminated data stream is input into the low-pass filter, and the cutoff frequency of the filter is set to remove noise higher than the frequency. The low-pass filter smoothes the signal and filters out high-frequency interference (such as electrical noise or high-frequency interference), and outputs a high-frequency denoised data stream. This data stream contains pure physiological indicator information after the high-frequency noise is removed, and finally a high-frequency denoised data stream of the patient's physiological indicators is obtained.

[0114] S23: performing physiological index baseline correction processing on the high-frequency denoised data stream of the patient's physiological index to obtain a physiological index baseline correction data stream of the patient;

[0115] In an embodiment of the present invention, a corresponding baseline drift component is extracted from a high-frequency denoised data stream of a patient's physiological indicators, usually by calculating a moving average of the data stream or using a regression method, and a corresponding baseline drift correction index is quantitatively calculated from the detected baseline drift component, and then subtracted from the original data stream to correct the baseline of the data stream. This correction process can use a baseline correction method to ensure that the physiological indicator data is analyzed at a preset baseline level, and ultimately obtain a baseline-corrected data stream of the patient's physiological indicators.

[0116] S24: performing time series synchronization processing on each physiological indicator in the patient physiological indicator baseline correction data stream to obtain the patient physiological indicator data stream in the same time series dimension;

[0117] In an embodiment of the present invention, each physiological indicator in the baseline correction data stream of the patient's physiological indicators obtained after the baseline correction is previously obtained is subjected to time series synchronization processing to extract timestamp data from different physiological indicator data streams, and align the time series of each data stream, so as to fill the time intervals between the data by using an interpolation method (such as linear interpolation or spline interpolation), thereby ensuring that all physiological indicators have valid data at the same time point, and finally obtaining the patient's physiological indicator data stream in the same time series dimension.

[0118] S25: Perform physiological feature vector conversion processing on the patient's physiological indicator data stream in the same time series dimension to obtain the patient's real-time physiological indicator change feature vector.

[0119] In an embodiment of the present invention, a corresponding feature analysis method is applied to the patient's physiological indicator data stream obtained in the same time series dimension after previous time series synchronization to analyze and extract key physiological characteristics, such as mean, standard deviation, frequency domain characteristics, etc., and by utilizing a feature vector generation method (such as principal component analysis PCA), the extracted feature data is converted into a physiological feature vector, which summarizes the change information of the patient's real-time physiological indicators, and finally obtains the patient's real-time physiological indicator change feature vector.

[0120] Furthermore, the physiological index baseline correction processing of the high-frequency noise-removing data stream of the patient's physiological index includes:

[0121] Performing frequency domain analysis on the high-frequency denoised data stream of the patient's physiological indicators to generate a frequency distribution spectrum of the patient's physiological indicator signals;

[0122] In an embodiment of the present invention, the high-frequency denoised data stream of the patient's physiological indicators obtained after denoising is applied to a Fourier transform tool. This process uses a discrete Fourier transform (DFT) algorithm to convert the time domain signal into a frequency domain signal. The frequency domain data after Fourier transform can be used to generate a frequency distribution map through a spectrum visualization tool. The spectrum map will display the amplitudes of different frequency components to facilitate the identification of the frequency distribution in the physiological signal, and finally generate a frequency distribution map of the patient's physiological indicator signal.

[0123] Preferably, based on the frequency distribution spectrum of the patient's physiological indicator signal, a physiological indicator baseline drift detection is performed on the corresponding patient's physiological indicator high-frequency noise-removed data stream to obtain the patient's physiological indicator signal distribution baseline drift trend data;

[0124] In an embodiment of the present invention, baseline drift detection is performed on the corresponding physiological indicators in the high-frequency denoised data stream of the patient's physiological indicators by combining the previously generated frequency distribution map of the physiological indicator signal, so as to identify the baseline drift component in the signal by analyzing the frequency distribution map, which is usually manifested as an abnormal component in the low-frequency region of the spectrum, and apply a baseline drift detection algorithm, such as a sliding average filter, to process the physiological indicator change signal in the map, so as to detect the baseline drift that can reflect the change trend of the patient's physiological indicators. This process uses statistical analysis tools to calculate the deviation trend of the signal, and finally obtains the baseline drift trend data of the patient's physiological indicator signal distribution.

[0125] Preferably, a baseline drift correction analysis is performed on the baseline drift trend data of the patient's physiological indicator signal distribution to obtain a baseline drift correction index of the patient's physiological indicator signal distribution;

[0126] In an embodiment of the present invention, the baseline drift trend data of the patient's physiological indicator signal distribution obtained by the previous analysis is fitted with a least square method or other regression analysis method, so as to fit a drift curve based on the baseline drift trend data, and the baseline drift part in the physiological indicator signal is stripped from the original signal to quantitatively calculate the corresponding drift correction index, for example Where I is the drift correction index, δ is the scaling factor, β0, β1 and β2 are the linear, first-order and quadratic coefficients in the baseline drift trend, β3 and β4 are the coefficients describing the periodic fluctuations in the baseline drift, corresponding to the sine and cosine terms respectively, ω is the angular frequency of the periodic component, γ is the exponential term scaling factor, ∈ is the exponential term growth rate, and exp is the exponential function. This correction index reflects the change of the actual signal under the condition of baseline drift, which can effectively guide the subsequent correction processing, and finally the baseline drift correction index of the patient's physiological indicator signal distribution is quantitatively calculated.

[0127] Preferably, based on the baseline drift correction index of the patient physiological indicator signal distribution, the corresponding patient physiological indicator high-frequency denoised data stream is subjected to data stream baseline correction adjustment processing to obtain the patient physiological indicator baseline correction data stream.

[0128] In an embodiment of the present invention, the baseline drift correction index of the patient physiological indicator signal distribution obtained by the previous quantitative calculation is used to adjust the data stream baseline correction of the corresponding patient physiological indicator high-frequency denoised data stream, so that the correction index is applied to the high-frequency denoised data stream, and the baseline of the data stream is adjusted using the correction index, so that the data stream is restored to a standard baseline state. The specific operation includes subtracting the correction index from the original data, or adjusting the deviation of the data stream through an inverse transformation to generate a baseline-corrected data stream. This step ensures that the patient's physiological indicator data stream can more accurately reflect the actual physiological state after baseline correction, and finally obtains the patient's physiological indicator baseline-corrected data stream.

[0129] Furthermore, the physiological characteristic vector conversion process of the patient's physiological indicator data stream in the same time series dimension includes:

[0130] Performing statistical analysis on the variability characteristics of the heart rate of patients with chronic cardiovascular disease in the patient physiological indicator data stream in the same time series dimension to obtain the heart rate variability characteristics of patients with chronic cardiovascular disease;

[0131] In an embodiment of the present invention, by performing statistical analysis on the variability characteristics of the heart rate of patients with chronic cardiovascular disease in the patient's physiological indicator data stream in the same time series dimension, heart rate data is extracted from the patient's physiological indicator data stream. These data need to be cleaned and preprocessed to eliminate noise and outliers, and by applying statistical analysis methods, such as time domain analysis, frequency domain analysis and nonlinear analysis, heart rate variability indicators such as standard deviation (SDNN), root mean square difference (RMSSD) and spectral power (LF / HF ratio) are calculated to obtain detailed characteristics of heart rate variability, and finally obtain the heart rate variability characteristics of patients with chronic cardiovascular disease.

[0132] Preferably, a waveform characteristic statistical analysis is performed on the ECG signal of a chronic cardiovascular disease patient in the patient physiological index data stream in the same time series dimension to obtain the waveform characteristics of the ECG signal of the chronic cardiovascular disease patient;

[0133] In an embodiment of the present invention, by performing statistical analysis on the signal waveform change characteristics of the ECG signal of a chronic cardiovascular disease patient in the patient's physiological indicator data stream under the same time series dimension, the ECG signal is extracted from the physiological indicator data stream, and denoising is performed to improve the signal quality. By applying waveform analysis techniques, such as QRS complex detection, ST segment analysis, and T wave detection, the main waveform features of the electrocardiogram are extracted, and specific software tools are used to extract features from the signal, and various waveform parameters, such as QRS width, ST segment offset, and T wave height, are calculated. These data are summarized as ECG signal waveform features, and finally the ECG signal waveform features of patients with chronic cardiovascular disease are obtained.

[0134] Preferably, the blood oxygen activity level characteristic analysis is performed on the blood oxygen saturation of the chronic cardiovascular disease patient in the patient physiological indicator data stream in the same time series dimension to obtain the blood oxygen activity level characteristic of the chronic cardiovascular disease patient;

[0135] In an embodiment of the present invention, a statistical analysis of the blood oxygen activity level of a chronic cardiovascular disease patient is performed on the blood oxygen saturation in the patient's physiological indicator data stream in the same time series dimension, so as to extract the blood oxygen saturation data from the patient's physiological indicator data stream, and after denoising and correction processing, the variability and fluctuation of the blood oxygen saturation are calculated by using signal processing methods such as moving average filtering and frequency domain analysis. The analysis results include the mean and standard deviation of the blood oxygen level and the time series changes of the blood oxygen activity. All the results are organized into blood oxygen activity level characteristics, and finally the blood oxygen activity level characteristics of the chronic cardiovascular disease patient are obtained.

[0136] Preferably, the blood pressure of chronic cardiovascular disease patients in the patient physiological indicator data stream under the same time series dimension is analyzed for abnormal blood pressure fluctuation characteristics to obtain abnormal blood pressure fluctuation characteristics of chronic cardiovascular disease patients;

[0137] In an embodiment of the present invention, a statistical analysis of abnormal fluctuations in the blood pressure of patients with chronic cardiovascular disease in a patient physiological indicator data stream under the same time series dimension is performed to extract blood pressure data from the data stream and preprocess it to remove interference and outliers. The abnormal fluctuations in blood pressure are evaluated through blood pressure fluctuation analysis techniques, such as statistical mean, standard deviation and fluctuation range calculation. At the same time, time series analysis methods are applied to identify and record abnormal fluctuation characteristics such as abnormal fluctuation patterns and frequencies of blood pressure, and finally the abnormal fluctuation characteristics of blood pressure in patients with chronic cardiovascular disease are obtained.

[0138] Preferably, the heart rate variability characteristics of chronic cardiovascular disease patients, the ECG signal waveform characteristics of chronic cardiovascular disease patients, the blood oxygen activity level characteristics of chronic cardiovascular disease patients and the abnormal blood pressure fluctuation characteristics of chronic cardiovascular disease patients are converted into physiological feature vectors to obtain the patient's real-time physiological indicator change feature vectors.

[0139] In an embodiment of the present invention, the heart rate variability characteristics of chronic cardiovascular disease patients, the ECG signal waveform characteristics of chronic cardiovascular disease patients, the blood oxygen activity level characteristics of chronic cardiovascular disease patients, and the abnormal blood pressure fluctuation characteristics of chronic cardiovascular disease patients obtained by previous statistical analysis are standardized to ensure that different features are compared at the same scale, and these features are integrated into a unified physiological feature vector by utilizing principal component analysis (PCA) or other dimensionality reduction techniques. This feature vector represents the changes in the patient's real-time physiological indicators and is updated in the real-time monitoring system to ultimately obtain the patient's real-time physiological indicator change feature vector.

[0140] Furthermore, the physiological indicator characteristic collapse abnormality quantification module includes the following functions:

[0141] Obtain physiological index data of patients with historical chronic cardiovascular cases;

[0142] In an embodiment of the present invention, physiological indicator data of patients with historical chronic cardiovascular cases are obtained through a medical database or recording system. These data include but are not limited to heart rate, ECG signal, blood oxygen saturation, and blood pressure. These data should be extracted from an electronic health record (EHR) system to ensure that the extracted data has time series characteristics and contains the patient's historical physiological measurement records. The data extraction tool must have the ability to process large amounts of data, such as an SQL database query tool, or use an API interface to export data from a health information system. The extracted records should include the patient's unique identifier and the corresponding timestamp, and ultimately the physiological indicator data of patients with historical chronic cardiovascular cases are obtained.

[0143] Preferably, the physiological index data of historical chronic cardiovascular patients are converted into consistent physiological characteristic vectors to obtain the physiological index change characteristic vectors of historical patients;

[0144] In an embodiment of the present invention, feature engineering technology is used to convert the acquired physiological indicator data of historical chronic cardiovascular patients into consistent feature vectors with previous physiological indicator real-time data streams, so as to realize the conversion of the physiological indicator data of historical chronic cardiovascular patients into corresponding physiological feature vectors (including historical feature data such as heart rate variability, ECG signal waveform, blood oxygen activity level and abnormal blood pressure fluctuations), and by applying principal component analysis (PCA) or other dimensionality reduction techniques, the high-dimensional physiological data is mapped to a low-dimensional feature space to obtain the corresponding feature vector, wherein the dimension of the feature vector should be selected according to the complexity of the data and the number of features, so as to effectively capture the changing characteristics of the patient's physiological data, and finally obtain the feature vector of the change of physiological indicators of the historical case patients.

[0145] Preferably, the feature vector space synchronization mapping process is performed on the feature vector of the change of the physiological index of the patient in the historical case and the feature vector of the change of the real-time physiological index of the patient, so as to obtain the feature vector of the physiological index of the historical medical record and the feature vector of the real-time physiological index in the same feature vector space dimension;

[0146] In an embodiment of the present invention, the physiological indicator feature vectors of historical cases and the real-time physiological indicator feature vectors are mapped to the same feature space by using feature vector space synchronous mapping to ensure that their dimensions are consistent with the feature space of historical case data, and by applying a spatial alignment algorithm, such as an iterative closest point (ICP) algorithm or a canonical correlation analysis (CCA), the historical feature vectors and the real-time feature vectors are mapped to the same feature space. This process ensures that the two sets of data are compared under the same feature dimension, thereby performing a consistency assessment, and finally obtaining the physiological indicator feature vectors of historical medical records and the real-time physiological indicator feature vectors under the same feature vector space dimension.

[0147] Preferably, based on the physiological indicator feature vectors of the historical medical records in the same feature vector space dimension, the corresponding real-time physiological indicator feature vectors are quantitatively calculated for the abnormal collapse of physiological indicator features to obtain abnormal collapse values ​​of the patient's real-time physiological indicator change features.

[0148] In an embodiment of the present invention, by calculating the feature space distance measurement value between each pair of corresponding physiological indicator features, a measurement method such as Euclidean distance and Manhattan distance can be used. The specific operation is to calculate the difference between each feature vector element of the historical medical record and the element at the corresponding position in the real-time physiological indicator feature vector, and square these differences, sum them up and then take the square root to obtain the feature space distance value. At the same time, by performing a benchmark mapping analysis of the feature vector space on the physiological indicator feature vectors of the historical medical record and the real-time physiological indicator feature vectors, a benchmark mapping matrix that maps them to the same feature space is calculated, and by combining the feature vector space benchmark mapping matrix obtained by the previous analysis, a differential analysis of the feature space reconstruction dimension of the real-time physiological indicator feature vector is performed to reconstruct the real-time physiological indicator feature vector through the benchmark mapping matrix. Convert to a new feature space, and then compare with the reconstruction of the feature vector in the original space to calculate the difference between the reconstructed and mapped feature vector and the original feature vector to obtain the corresponding vector space reconstruction dimension difference statistics, and construct a suitable calculation formula by combining the feature space distance value and the vector space reconstruction dimension difference statistics to perform quantitative calculation of the collapse abnormality of the physiological indicator feature vector for the corresponding real-time physiological indicator feature vector. This formula can be in the form of a weighted sum or a weighted difference to quantitatively calculate the collapse abnormality value of the corresponding physiological indicator feature vector. This collapse abnormality value reflects whether the change of the real-time physiological indicator deviates significantly from the normal range in the historical data, thereby helping to detect potential physiological abnormalities, and finally obtaining the collapse abnormality value of the patient's real-time physiological indicator change characteristics.

[0149] Furthermore, the physiological indicator characteristic collapse abnormality quantitative calculation of the corresponding real-time physiological indicator characteristic vector based on the historical medical record physiological indicator characteristic vector in the same characteristic vector space dimension includes:

[0150] Performing feature space distance measurement calculation between each corresponding physiological indicator feature in the historical medical record physiological indicator feature vector and the real-time physiological indicator feature vector under the same feature vector space dimension, so as to obtain a feature space distance measurement value between each historical physiological indicator feature vector and the corresponding real-time physiological indicator feature vector;

[0151] In an embodiment of the present invention, by collecting physiological indicator feature vectors of historical medical records and real-time physiological indicator feature vectors, these feature vectors are analyzed in the same dimension, and by calculating the feature space distance measurement value between each pair of corresponding physiological indicator features, Euclidean distance, Manhattan distance and other measurement methods can be used. The specific operation is to calculate the difference between each feature vector element of the historical medical record and the element at the corresponding position in the real-time physiological indicator feature vector, and then square these differences, sum them and take the square root to obtain the feature space distance value. By writing an algorithm script and using Python libraries such as NumPy for numerical calculation and processing, it is ensured that the distance between each historical feature vector and its corresponding real-time feature vector is accurately measured, and finally the feature space distance measurement value between each historical physiological indicator feature vector and the corresponding real-time physiological indicator feature vector is obtained.

[0152] Preferably, a feature vector space reference mapping analysis is performed on the historical medical record physiological indicator feature vectors and the real-time physiological indicator feature vectors in the same feature vector space dimension to obtain a physiological indicator feature vector space reference mapping matrix;

[0153] In an embodiment of the present invention, a benchmark mapping analysis of the feature vector space is performed on the historical medical record physiological indicator feature vectors and the real-time physiological indicator feature vectors. This step can be achieved by the principal component analysis (PCA) method to merge the historical medical record and real-time physiological indicator feature vectors into one data set, and then apply the PCA algorithm to determine the benchmark mapping matrix of the feature space. In the specific implementation process, the PCA module in the Scikit-learn library of Python is used to train the corresponding physiological indicator feature vectors to calculate the benchmark mapping matrix that maps them to the same feature space. This matrix will serve as the basis for subsequent analysis to map the feature vectors to the new feature space, and finally obtain the physiological indicator feature vector space benchmark mapping matrix.

[0154] Preferably, based on the physiological indicator feature vector space reference mapping matrix, feature space reconstruction dimension difference analysis is performed on the real-time physiological indicator feature vectors in the same feature vector space dimension to obtain the patient's real-time physiological indicator feature vector space reconstruction dimension difference value;

[0155] In an embodiment of the present invention, a differential analysis of the feature space reconstruction dimension is performed on the real-time physiological indicator feature vector by combining the feature vector space reference mapping matrix obtained by previous analysis, so that the real-time physiological indicator feature vector is converted to a new feature space through the reference mapping matrix, and then compared with the reconstruction of the feature vector in the original space. The dimension difference value can be obtained by calculating the difference between the reconstructed mapped feature vector and the original feature vector. The specific method includes performing element-level difference between the reconstructed mapped vector and the original vector to calculate the average or other statistics of the absolute values ​​of these differences, and using tools such as Python's SciPy library to perform these mathematical operations, and finally obtaining the spatial reconstruction dimension difference value of the patient's real-time physiological indicator feature vector.

[0156] Preferably, based on the feature space distance measurement value between each historical physiological indicator feature vector and the corresponding real-time physiological indicator feature vector and the spatial reconstruction dimension difference value of the patient's real-time physiological indicator feature vector, the physiological indicator feature collapse abnormality calculation formula is used to perform a quantitative calculation of the physiological indicator feature collapse abnormality on the corresponding real-time physiological indicator feature vector to obtain the patient's real-time physiological indicator change feature collapse abnormality value.

[0157] In an embodiment of the present invention, a suitable physiological indicator characteristic collapse abnormality calculation formula is formed by combining the initial integration time variable, the end integration time variable, the integration time variable parameter, the quantity parameter of the physiological indicator characteristic vector, the historical physiological indicator characteristic vector and the real-time physiological indicator characteristic vector in the same characteristic vector space dimension, the characteristic space distance measurement paradigm, the distance standard deviation between the physiological indicator characteristic vectors, the characteristic vector space distance weighting coefficient, the patient's real-time physiological indicator characteristic vector space reconstruction dimension difference value, the space reconstruction dimension difference weighting coefficient, the physiological indicator characteristic vector space reference mapping matrix, the matrix transposition symbol, the real-time physiological indicator characteristic vector in the original space dimension and related parameters to perform a quantitative calculation of the physiological indicator characteristic collapse abnormality on the corresponding real-time physiological indicator characteristic vector. This formula can be in the form of a weighted sum or a weighted difference to quantitatively calculate the collapse abnormality value of the corresponding physiological indicator characteristic vector, and finally obtain the patient's real-time physiological indicator change characteristic collapse abnormality value.

[0158] Furthermore, the calculation formula for abnormal collapse of physiological index characteristics is specifically:

[0159]

[0160] ΔD(t)=U T *(X r,i (t)―X l,i (t));

[0161] In the formula, E cis the abnormal value of the collapse of the characteristic of the patient's real-time physiological index change, t1 is the initial time variable of integration, t2 is the end time variable of integration, t is the integral time variable parameter, n is the total number of physiological index feature vectors, i is the item measurement parameter of the physiological index feature vector, X s,i (t) is the i-th historical physiological index feature vector at time t in the same feature vector space dimension, X l,i (t) is the i-th real-time physiological index feature vector corresponding to time t in the same feature vector space dimension, ‖·‖ 2 is the feature space distance measurement paradigm, σ i is the standard deviation of the distance between the ith physiological index feature vectors, α is the feature vector space distance weighting coefficient, ΔD(t) is the difference value of the spatial reconstruction dimension of the patient's real-time physiological index feature vector at time t, β is the spatial reconstruction dimension difference weighting coefficient, U is the physiological index feature vector space reference mapping matrix, T is the matrix transposition symbol, X r,i (t) is the i-th real-time physiological index feature vector at time t in the original spatial dimension, and η is the correction coefficient of the abnormal value of the collapse of the patient's real-time physiological index change feature.

[0162] The present invention obtains a physiological indicator characteristic collapse abnormality calculation formula by using a specific mathematical model and after verification, which is used to perform physiological indicator characteristic collapse abnormality quantitative calculation on the corresponding real-time physiological indicator characteristic vector. The physiological indicator characteristic collapse abnormality calculation formula calculates the characteristic space distance measurement value between the historical physiological indicator characteristic vector and the real-time physiological indicator characteristic vector, which can quantify the change of each physiological indicator at the historical and current time points. This measurement provides the difference between the real-time data and the historical data, and helps to identify potential abnormal changes. The feature space reconstruction dimension difference analysis is performed through the reference mapping matrix, which provides the difference between the real-time physiological data and the historical data in the reconstructed space. This analysis helps to discover the changes in the dimension and capture the existing abnormal patterns or trends. The weighting coefficient in the formula allows different features to be given different importance when calculating abnormal values. This weighting mechanism helps to more accurately reflect which features occupy an important position in the change of physiological indicators, and avoids the excessive influence of certain features on the calculation of abnormal values. By incorporating the difference between the real-time physiological indicators and the historical data into the calculation, the formula takes into account the dynamic changes in time, which is particularly important for capturing gradual changes or sudden abnormalities. At the same time, the formula uses an integral form to accumulate the feature space distance measure and the reconstructed dimension difference value. This method allows a comprehensive evaluation of abnormal conditions within a certain time interval, providing a dynamic perspective to monitor changes in the patient's physiological state. In addition, by using the correction coefficient to adjust the calculation results, it can take into account measurement errors or other external factors. This correction makes the calculation of abnormal values ​​more reliable and closer to the actual situation. By quantifying the characteristic collapse abnormal values ​​of changes in physiological indicators, doctors can formulate or adjust plans based on these indicators. This quantitative result provides a clear and numerical basis, which helps to make more accurate decisions in clinical practice. This formula and its calculation steps combine a variety of methods and techniques to conduct in-depth analysis and evaluation of physiological indicator data. It not only improves the accuracy and reliability of abnormality detection, but also provides doctors with a comprehensive tool to understand and respond to changes in patients' physiological states. In summary, this formula can fully consider the characteristic collapse abnormal values ​​E of real-time changes in patients' physiological indicators. c , integration initial time variable t1, integration termination time variable t2, integration time variable parameter t, the total number of physiological index feature vectors n, the item measurement parameter i of the physiological index feature vector, the i-th historical physiological index feature vector X at time t in the same feature vector space dimension s,i (t), the i-th real-time physiological index feature vector X corresponding to time t in the same feature vector space dimension l,i (t), feature space distance measurement paradigm ‖·‖ 2 , the standard deviation of the distance between the i-th physiological index feature vectors σ i, feature vector space distance weighting coefficient α, the patient's real-time physiological index feature vector space reconstruction dimension difference value ΔD(t) at time t, spatial reconstruction dimension difference weighting coefficient β, physiological index feature vector space reference mapping matrix U, matrix transposition symbol T, the i-th real-time physiological index feature vector X at time t in the original space dimension r,i (t), the correction coefficient η of the abnormal value of the collapse of the patient's real-time physiological index change characteristic, where the item measurement parameter i of the physiological index feature vector, the integral time variable parameter t, and the i-th real-time physiological index feature vector X corresponding to time t in the same feature vector space dimension l,i (t), physiological index feature vector space reference mapping matrix U, matrix transposition symbol T and the i-th real-time physiological index feature vector X at time t in the original space dimension r,i (t) constitutes a functional relationship U of the patient's real-time physiological index feature vector space reconstruction dimension difference value ΔD(t) at time t T *(X r,i (t)―X l,i (t)) according to the patient's real-time physiological index change characteristics, collapse the abnormal value E c The correlation between the above parameters constitutes a functional relationship This formula can realize the quantitative calculation process of the physiological indicator characteristic collapse abnormality of the corresponding real-time physiological indicator characteristic vector. At the same time, by introducing the correction coefficient η of the abnormal collapse value of the patient's real-time physiological indicator change characteristic, it can be adjusted according to the error situation occurring in the calculation process, thereby improving the accuracy and applicability of the physiological indicator characteristic collapse abnormality calculation formula.

[0163] Furthermore, the patient physiological index abnormal response warning module includes the following functions:

[0164] Compare and judge the abnormal collapse value of the patient's real-time physiological indicator change characteristic according to the preset abnormal collapse threshold of the physiological indicator characteristic. When the abnormal collapse value of the patient's real-time physiological indicator change characteristic is greater than or equal to the preset abnormal collapse threshold of the physiological indicator characteristic, perform abnormal feature mapping processing on the corresponding patient's real-time physiological indicator change characteristic vector to generate an abnormal feature mapping map of the patient's physiological indicator; when the abnormal collapse value of the patient's real-time physiological indicator change characteristic is less than the preset abnormal collapse threshold of the physiological indicator characteristic, no processing is performed;

[0165] In an embodiment of the present invention, by using a preset physiological index characteristic collapse abnormal threshold, such as the normal fluctuation range of physiological data such as heart rate, respiratory rate, body temperature, etc., during the real-time monitoring process, the patient's physiological index change characteristic data is collected, and the corresponding patient's real-time physiological index change characteristic collapse abnormal value is calculated. If the real-time collected physiological index change characteristic collapse abnormal value exceeds or equals the preset abnormal threshold, the corresponding patient's real-time physiological index characteristic vector is subjected to abnormal feature mapping processing. The data used in the mapping process may include a specific abnormal identification algorithm, such as k-means clustering, support vector machine (SVM), etc., to convert the corresponding physiological index change characteristic vector into an abnormal feature mapping spectrum. During the spectrum generation process, data visualization tools, such as MATLAB's drawing function, or special abnormality detection software, can be used to intuitively display the abnormal features as a spectrum, thereby generating a patient's physiological index abnormal feature mapping spectrum. If the real-time characteristic collapse abnormal value does not reach the abnormal threshold, no processing is performed to save computing resources and avoid false alarms.

[0166] Preferably, based on the abnormal characteristic map of the patient's physiological indicator, abnormal physiological indicator response processing is performed on the corresponding patient's physiological indicator in the patient's real-time physiological indicator change characteristic vector to generate an abnormal physiological indicator change response signal of the patient;

[0167] In an embodiment of the present invention, abnormal response processing of physiological indicators is performed on the corresponding patient physiological indicators in the characteristic vector of the patient's real-time physiological indicator changes by combining the abnormal characteristic mapping map of the patient's physiological indicators drawn previously. The analysis process of the map can use data mining technology, such as anomaly detection algorithm, to identify specific abnormal physiological indicators by marking the abnormal areas in the map, and analyze the relationship between the abnormal characteristic map and the real-time physiological data by applying a response processing algorithm, such as dynamic time warping (DTW) or other matching algorithms. The generated abnormal response signal includes the real-time value of the corresponding abnormal indicator, the abnormal level and other information. This signal is used for subsequent early warning processing to accurately reflect it for further intervention, and finally respond to generate an abnormal response signal of the patient's physiological indicator change.

[0168] Preferably, the abnormal response signal of the patient's physiological indicator change is sent through the wireless transmission module and acts on the early warning controller in the smart wearable device to perform abnormal early warning control processing, generate the patient's physiological indicator abnormal early warning prompt control instruction, so as to execute the corresponding patient's physiological indicator abnormal early warning prompt work.

[0169] In an embodiment of the present invention, the generated abnormal response signal of the patient's physiological indicator change is wirelessly transmitted by using the built-in wireless transmission module of the smart wearable device. The wireless transmission module can be Bluetooth, Wi-Fi or other wireless communication technologies. The abnormal signal is sent to the early warning controller in the smart wearable device. After receiving the signal, the early warning controller processes it through a preset control logic, such as setting threshold judgment, abnormal level classification, etc., so as to respond and generate corresponding patient physiological indicator abnormal early warning prompt control instructions, including multiple reminder methods such as vision, hearing or vibration, and send warnings to users or medical staff in real time through smart wearable devices. For example, the controller will activate the alarm sound on the smart wearable device or the red warning icon on the display screen, so as to deal with the abnormal situation in time and ensure that the patient can get quick intervention, thereby executing the corresponding patient physiological indicator abnormal early warning prompt work.

[0170] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0171] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A data monitoring system for chronic cardiovascular disease, characterized in that: Includes the following modules: The patient physiological index data stream acquisition module is used to collect the real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease through smart wearable devices in real time, and perform stream data fusion processing on the real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease to obtain the real-time data stream of the patient's physiological index; the real-time data stream of the patient's physiological index is sent and transmitted to the data processing unit through the wireless transmission module; The physiological index change characteristic vector conversion module is used to perform physiological index baseline correction processing on the patient's physiological index real-time data stream through the data processing unit to obtain the patient's physiological index baseline correction data stream; perform time series synchronization processing on each physiological index in the patient's physiological index baseline correction data stream to obtain the patient's physiological index data stream under the same time series dimension; perform physiological characteristic vector conversion processing on the patient's physiological index data stream under the same time series dimension to obtain the patient's real-time physiological index change characteristic vector; The physiological index characteristic collapse abnormality quantification module is used to obtain the physiological index data of patients with historical chronic cardiovascular cases, and perform consistent physiological characteristic vector conversion on the physiological index data of patients with historical chronic cardiovascular cases to obtain the physiological index change characteristic vector of the patients with historical cases; based on the physiological index change characteristic vector of the patients with historical cases, the physiological index characteristic collapse abnormality quantification calculation is performed on the patient's real-time physiological index change characteristic vector to obtain the patient's real-time physiological index change characteristic collapse abnormality value; The patient physiological indicator abnormal response warning module is used to perform physiological indicator abnormal response processing on the corresponding patient physiological indicators in the patient's real-time physiological indicator change feature vector based on the abnormal value of the patient's real-time physiological indicator change characteristic collapse, so as to generate a patient physiological indicator change abnormal response signal; the patient physiological indicator change abnormal response signal is applied to the smart wearable device to execute the corresponding patient physiological indicator abnormal warning prompt work.

2. The data monitoring system for chronic cardiovascular disease according to claim 1, characterized in that: The patient's physiological index data stream acquisition module Features include: Collect real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease through smart wearable devices; The real-time physiological index data of heart rate, ECG signal, blood oxygen saturation and blood pressure of patients with chronic cardiovascular disease are subjected to stream data mapping and conversion through the stream data processing architecture to generate a sub-stream of heart rate data of patients with chronic cardiovascular disease, a sub-stream of ECG signal data of patients with chronic cardiovascular disease, a sub-stream of blood oxygen saturation data of patients with chronic cardiovascular disease and a sub-stream of blood pressure data of patients with chronic cardiovascular disease; Performing stream data fusion processing on the heart rate data substream, the ECG signal data substream, the blood oxygen saturation data substream and the blood pressure data substream of the chronic cardiovascular disease patients to obtain the real-time data stream of the patients' physiological indicators; The real-time data stream of the patient's physiological indicators is sent and transmitted to the data processing unit through the wireless transmission module.

3. The data monitoring system for chronic cardiovascular disease according to claim 2, characterized in that: The stream data fusion processing of the heart rate data substream of the chronic cardiovascular disease patient, the ECG signal data substream of the chronic cardiovascular disease patient, the blood oxygen saturation data substream of the chronic cardiovascular disease patient, and the blood pressure data substream of the chronic cardiovascular disease patient comprises: Performing data stream hierarchical extraction processing on the heart rate data substream of chronic cardiovascular disease patients, the ECG signal data substream of chronic cardiovascular disease patients, the blood oxygen saturation data substream of chronic cardiovascular disease patients, and the blood pressure data substream of chronic cardiovascular disease patients to obtain hierarchical physiological indicator data streams of chronic cardiovascular disease patients; Performing statistical analysis on the time series characteristics of the data stream of stratified physiological indicators of patients with chronic cardiovascular disease to obtain a data set of time series characteristics of stratified physiological indicators of patients with chronic cardiovascular disease; Performing feature smoothing and enhancement processing on the time series features of each physiological indicator data substream in the time series feature data set of stratified physiological indicators of patients with chronic cardiovascular disease, so as to obtain the enhanced time series feature data set of stratified physiological indicators of patients; The temporal correlation mining analysis is performed on the temporal correlation between different physiological indicator temporal characteristics in the patient-stratified physiological indicator enhanced temporal characteristic data set to obtain the temporal correlation relationship between different physiological indicator temporal characteristics of chronic cardiovascular disease patients; based on the temporal correlation relationship between different physiological indicator temporal characteristics of chronic cardiovascular disease patients, the correlation feature map is connected and constructed between different physiological indicator temporal characteristics in the patient-stratified physiological indicator enhanced temporal characteristic data set to generate a dynamic correlation feature map of patient physiological indicators; Based on the dynamic correlation feature map of patients' physiological indicators, the heart rate data substream of chronic cardiovascular disease patients, the ECG signal data substream of chronic cardiovascular disease patients, the blood oxygen saturation data substream of chronic cardiovascular disease patients and the blood pressure data substream of chronic cardiovascular disease patients are fused to obtain the real-time data stream of patients' physiological indicators.

4. The data monitoring system for chronic cardiovascular disease according to claim 1, characterized in that: The physiological index change feature vector conversion module includes the following functions: Performing low-frequency interference elimination processing on the real-time data stream of the patient's physiological indicators through a high-pass filter in the data processing unit to obtain a low-frequency interference eliminated data stream of the patient's physiological indicators; Performing high-frequency noise elimination processing on the low-frequency interference elimination data stream of the patient's physiological indicators through a low-pass filter in the data processing unit to obtain a high-frequency noise elimination data stream of the patient's physiological indicators; Performing physiological index baseline correction processing on the high-frequency denoised data stream of the patient's physiological index to obtain a physiological index baseline correction data stream of the patient; Performing time series synchronization processing on each physiological indicator in the patient's physiological indicator baseline correction data stream to obtain the patient's physiological indicator data stream in the same time series dimension; The physiological indicator data stream of the patient in the same time series dimension is converted into a physiological feature vector to obtain a feature vector of the patient's real-time physiological indicator change.

5. The data monitoring system for chronic cardiovascular disease according to claim 4, characterized in that: The physiological index baseline correction processing of the high-frequency noise-removing data stream of the patient's physiological index comprises: Performing frequency domain analysis on the high-frequency denoised data stream of the patient's physiological indicators to generate a frequency distribution spectrum of the patient's physiological indicator signals; Based on the frequency distribution spectrum of the patient's physiological indicator signal, a physiological indicator baseline drift detection is performed on the corresponding patient's physiological indicator high-frequency noise-removing data stream to obtain the patient's physiological indicator signal distribution baseline drift trend data; Performing baseline drift correction analysis on the baseline drift trend data of the patient's physiological indicator signal distribution to obtain the baseline drift correction index of the patient's physiological indicator signal distribution; Based on the baseline drift correction index of the patient's physiological indicator signal distribution, the corresponding patient's physiological indicator high-frequency denoised data stream is subjected to data stream baseline correction and adjustment processing to obtain the patient's physiological indicator baseline correction data stream.

6. The data monitoring system for chronic cardiovascular disease according to claim 4, characterized in that: The physiological characteristic vector conversion process of the patient's physiological indicator data stream in the same time series dimension includes: Performing statistical analysis on the variability characteristics of the heart rate of patients with chronic cardiovascular disease in the patient physiological indicator data stream in the same time series dimension to obtain the heart rate variability characteristics of patients with chronic cardiovascular disease; Performing signal waveform feature statistical analysis on ECG signals of patients with chronic cardiovascular disease in the patient physiological indicator data stream under the same time series dimension to obtain ECG signal waveform features of patients with chronic cardiovascular disease; Perform blood oxygen activity level feature analysis on the blood oxygen saturation of chronic cardiovascular disease patients in the patient physiological indicator data stream in the same time series dimension to obtain the blood oxygen activity level feature of chronic cardiovascular disease patients; The abnormal fluctuation characteristics of blood pressure of chronic cardiovascular disease patients in the patient physiological indicator data stream under the same time series dimension are analyzed to obtain the abnormal fluctuation characteristics of blood pressure of chronic cardiovascular disease patients; The heart rate variability characteristics of chronic cardiovascular disease patients, the ECG signal waveform characteristics of chronic cardiovascular disease patients, the blood oxygen activity level characteristics of chronic cardiovascular disease patients, and the abnormal blood pressure fluctuation characteristics of chronic cardiovascular disease patients are converted into physiological feature vectors to obtain the feature vectors of real-time physiological index changes of patients.

7. The data monitoring system for chronic cardiovascular disease according to claim 1, characterized in that: The physiological indicator characteristic collapse abnormality quantification module includes the following functions: Obtain physiological index data of patients with historical chronic cardiovascular cases; Perform consistent physiological characteristic vector conversion on the physiological indicator data of historical chronic cardiovascular patients to obtain the characteristic vector of physiological indicator changes of historical patients; Performing feature vector space synchronization mapping processing on the feature vectors of changes in physiological indicators of historical cases and the feature vectors of changes in real-time physiological indicators of patients, so as to obtain the feature vectors of physiological indicators of historical medical records and the feature vectors of real-time physiological indicators in the same feature vector space dimension; Based on the physiological indicator feature vectors of historical medical records in the same feature vector space dimension, the corresponding real-time physiological indicator feature vectors are quantitatively calculated for the abnormal collapse of physiological indicator features to obtain the abnormal collapse value of the patient's real-time physiological indicator change features.

8. The data monitoring system for chronic cardiovascular disease according to claim 7, characterized in that: The method of performing physiological indicator characteristic collapse abnormality quantitative calculation on the corresponding real-time physiological indicator characteristic vector based on the historical medical record physiological indicator characteristic vector in the same characteristic vector space dimension comprises: Performing feature space distance measurement calculation between each corresponding physiological indicator feature in the historical medical record physiological indicator feature vector and the real-time physiological indicator feature vector under the same feature vector space dimension, so as to obtain a feature space distance measurement value between each historical physiological indicator feature vector and the corresponding real-time physiological indicator feature vector; Performing feature vector space benchmark mapping analysis on the feature vectors of historical medical records physiological indicators and the feature vectors of real-time physiological indicators in the same feature vector space dimension to obtain a physiological indicator feature vector space benchmark mapping matrix; Based on the physiological index characteristic vector space reference mapping matrix, the characteristic space reconstruction dimension difference analysis is performed on the real-time physiological index characteristic vectors in the same characteristic vector space dimension to obtain the patient's real-time physiological index characteristic vector space reconstruction dimension difference value; Based on the feature space distance measurement value between each historical physiological indicator feature vector and the corresponding real-time physiological indicator feature vector and the difference value of the spatial reconstruction dimension of the patient's real-time physiological indicator feature vector, the physiological indicator feature collapse abnormality calculation formula is used to quantitatively calculate the physiological indicator feature collapse abnormality of the corresponding real-time physiological indicator feature vector to obtain the patient's real-time physiological indicator change feature collapse abnormality value.

9. The data monitoring system for chronic cardiovascular disease according to claim 8, characterized in that: The specific calculation formula for abnormal collapse of physiological index characteristics is: ΔD(t)=U T *(X r，i (t)-X l，i (t)); In the formula, E c is the abnormal value of the collapse of the characteristic of the patient's real-time physiological index change, t1 is the initial time variable of integration, t2 is the end time variable of integration, t is the integral time variable parameter, n is the total number of physiological index feature vectors, i is the item measurement parameter of the physiological index feature vector, X s，i (t) is the i-th historical physiological index feature vector at time t in the same feature vector space dimension, X l，i (t) is the i-th real-time physiological index feature vector corresponding to time t in the same feature vector space dimension, ||·|| 2 is the feature space distance measurement paradigm, σ i is the standard deviation of the distance between the ith physiological index feature vectors, α is the feature vector space distance weighting coefficient, ΔD(t) is the difference value of the spatial reconstruction dimension of the patient's real-time physiological index feature vector at time t, β is the spatial reconstruction dimension difference weighting coefficient, U is the physiological index feature vector space reference mapping matrix, T is the matrix transposition symbol, X r，i (t) is the i-th real-time physiological index feature vector at time t in the original spatial dimension, and η is the correction coefficient of the abnormal value of the collapse of the patient's real-time physiological index change feature.

10. The data monitoring system for chronic cardiovascular disease according to claim 1, characterized in that: The patient's physiological index abnormal response warning module Features include: According to the preset abnormal collapse threshold of physiological indicator characteristics, the abnormal collapse value of the patient's real-time physiological indicator change characteristics is compared and judged. When the abnormal collapse value of the patient's real-time physiological indicator change characteristics is greater than or equal to the preset abnormal collapse threshold of physiological indicator characteristics, the corresponding patient's real-time physiological indicator change characteristic vector is subjected to abnormal characteristic mapping processing to generate an abnormal characteristic mapping map of the patient's physiological indicators; When the abnormal collapse value of the patient's real-time physiological indicator change characteristic is less than the preset abnormal collapse threshold of the physiological indicator characteristic, no processing is performed; Based on the abnormal characteristic mapping graph of the patient's physiological indicators, abnormal physiological indicator response processing is performed on the corresponding patient's physiological indicators in the patient's real-time physiological indicator change characteristic vector to generate an abnormal physiological indicator change response signal of the patient; The abnormal response signal of the patient's physiological index change is sent through the wireless transmission module and acts on the early warning controller in the smart wearable device to perform abnormal early warning control processing, generate the patient's physiological index abnormal early warning prompt control instruction, so as to execute the corresponding patient's physiological index abnormal early warning prompt work.

Images