Blood pressure prediction method and device, electronic equipment and computer readable storage medium

Abstract

The embodiment of the invention provides a blood pressure prediction method and device, electronic equipment and a computer readable storage medium, and relates to the technical field of computers. The method comprises the following steps: acquiring first physiological information and second physiological information of a target object in a preset time window; the first physiological information is information collected by a pulse pressure sensor, and the second physiological information is information collected by a photoelectric volume pulse wave map sensor; based on the plurality of evaluation indexes, selecting physiological information meeting a preset evaluation standard from the first physiological information and the second physiological information as target physiological information; inputting the target physiological information into a prediction model to obtain a prediction result output by the prediction model; the prediction result comprises systolic pressure and diastolic pressure of the target object; the prediction model is trained according to a training sample and a corresponding training label, and the training label is systolic pressure and diastolic pressure corresponding to the training sample; according to the embodiment of the invention, the problem of low blood pressure prediction accuracy under different prediction conditions is solved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and more specifically, to a blood pressure prediction method, device, electronic device, and computer-readable storage medium. Background Art

[0002] Cardiovascular health engineering has become an important field, driving innovations aimed at improving personal health monitoring and management. High blood pressure is a major risk factor for a variety of life-threatening conditions, including stroke, heart disease, diabetes, and even complications of certain cancers. Monitoring blood pressure is essential for early detection and control of high blood pressure.

[0003] However, the current blood pressure estimation method is easily affected by various factors such as the physiological characteristics of different users and the environment they are in, which may lead to inconsistency between the predicted blood pressure and the actual blood pressure. Therefore, the current blood pressure estimation method has the problem of low blood pressure prediction accuracy under different prediction conditions. Summary of the invention

[0004] The embodiments of the present application provide a blood pressure prediction method, device, electronic device and computer-readable storage medium, which are used to solve the technical problem of low blood pressure prediction accuracy under different measurement conditions.

[0005] According to a first aspect of an embodiment of the present application, a blood pressure prediction method is provided, the method comprising: obtaining first physiological information and second physiological information of a target object within a preset time window; the first physiological information is information collected by a pulse pressure sensor, and the second physiological information is information collected by a photoplethysmogram sensor; Based on the multiple evaluation indicators, selecting physiological information that meets the preset evaluation criteria from the first physiological information and the second physiological information as the target physiological information; Inputting the target physiological information into the prediction model to obtain the prediction results output by the prediction model; the prediction results include: systolic blood pressure and diastolic blood pressure of the target object; The prediction model is trained based on multiple training samples and corresponding training labels, the multiple training samples come from the first sample set and the second sample set, and the training labels are systolic pressure and diastolic pressure corresponding to the training samples; The first sample set is composed of first physiological information of the sample subject, and the second sample set is composed of second physiological information of the sample subject.

[0006] In a possible implementation, the prediction model includes a first sub-model and a second sub-model, and the first sub-model is trained according to training samples in the first sample set and corresponding training labels; The second sub-model is trained based on the training samples in the second sample set and the corresponding training labels; If the target physiological information is the first physiological information, the target physiological information is input into the first sub-model; If the target physiological information is the second physiological information, the target physiological information is input into the second sub-model.

[0007] In another possible implementation, the first sub-model includes a first grandchild model and a second grandchild model, and the second sub-model includes a third grandchild model and a fourth grandchild model; The first submodel is trained based on the training samples in the first sample set and the corresponding first training labels, and the first training labels are the systolic blood pressure of the training samples; The second Sun model is trained based on the training samples in the first sample set and the corresponding second training labels, and the second training labels are the diastolic pressure of the training samples; The third Sun model is trained based on the training samples in the second sample set and the corresponding third training labels, where the third training labels are the systolic blood pressure of the training samples; The fourth grandson model is trained according to the fourth training label corresponding to the first level of the training sample in the second sample set, and the fourth training label is the diastolic pressure of the training sample.

[0008] In another possible implementation, the prediction model includes: a feature extraction layer, a long short-term memory (LSTM) layer, and a fully connected layer; inputting the target physiological information into the feature extraction layer to obtain a feature vector sequence of the target physiological information output by the feature extraction layer; The feature vector sequence is input into the LSTM layer to obtain the time series characteristics of the target physiological information output by the LSTM layer; the time series characteristics are used to reflect the information characteristics of the target physiological information changing over time; Input the time series features into the fully connected layer to obtain the prediction results output by the fully connected layer.

[0009] In yet another possible implementation, a plurality of evaluation indicators and a weight of each evaluation indicator are determined; respectively calculating a first index value of each evaluation index of the first physiological information and a second index value of each evaluation index of the second physiological information; determining a first indicator score of the first physiological information based on first indicator values ​​and corresponding weights of each evaluation indicator of the first physiological information; determining a second indicator score of the second physiological information based on the second indicator value and the corresponding weight of each evaluation indicator of the second physiological information; From the first indicator score and the second indicator score, select the physiological information with a higher indicator score as the target physiological information; The evaluation indicators include at least one of the following: Signal-to-noise ratio of physiological information; Stability of physiological information; Continuity of physiological information.

[0010] In another possible implementation, after obtaining the prediction result output by the prediction model, if the prediction result meets the alarm condition, an alarm message is generated; the alarm message is used to indicate that the systolic blood pressure and / or diastolic blood pressure of the target object is abnormal; Display warning information in a preset manner.

[0011] In yet another possible implementation, the prediction result is sent to the bound terminal; The bound terminal includes at least one of the following: The target subject's medical service provider's terminal; Terminals for the target’s family members; The terminal of the caretaker of the target object.

[0012] According to a second aspect of an embodiment of the present application, a blood pressure prediction device is provided, the device comprising: An acquisition module, used to acquire first physiological information and second physiological information of a target object within a preset time window; the first physiological information is information acquired by a pulse pressure sensor, and the second physiological information is information acquired by a photoplethysmogram sensor; A selection module, configured to select, based on a plurality of evaluation indicators, physiological information that meets a preset evaluation standard from the first physiological information and the second physiological information as target physiological information; An input module is used to input the target physiological information into the prediction model to obtain the prediction results output by the prediction model; the prediction results include: systolic blood pressure and diastolic blood pressure of the target object; The prediction model is trained based on multiple training samples and corresponding training labels, the multiple training samples come from the first sample set and the second sample set, and the training labels are systolic pressure and diastolic pressure corresponding to the training samples; The first sample set is composed of first physiological information of the sample subject, and the second sample set is composed of second physiological information of the sample subject.

[0013] According to a third aspect of an embodiment of the present application, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory, and when the processor executes the program, the steps of the method provided in the first aspect are implemented.

[0014] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method provided in the first aspect are implemented.

[0015] According to the fifth aspect of the embodiment of the present application, a computer program product is provided, which includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. When a processor of a computer device reads the computer instructions from the computer-readable storage medium, the processor executes the computer instructions, so that the computer device executes the steps of implementing the method provided in the first aspect.

[0016] The beneficial effects of the technical solution provided by the embodiment of the present application are: The blood pressure prediction method provided in the embodiment of the present application obtains the first physiological information collected by the pulse pressure sensor and the second physiological information collected by the photoplethysmogram sensor within a preset time window, and then selects the physiological information that meets the preset evaluation criteria from the first physiological information and the second physiological information based on multiple evaluation indicators as the target physiological information, inputs the target physiological information into the prediction model, and obtains the systolic pressure and diastolic pressure of the target object output by the prediction model. Before inputting the collected physiological information into the prediction model, first selects the physiological information with higher quality from the first physiological information and the second physiological information based on multiple evaluation indicators as the target physiological information and inputs it into the prediction model, so as to ensure that the physiological information input into the prediction model is accurate and effective, and ensure that high-quality data is used for blood pressure prediction, thereby avoiding the influence of factors such as the current physiological characteristics of the user and the environment in which the user is located on the quality of the collected physiological signals, thereby affecting the final prediction result, and the prediction model used to predict blood pressure is trained based on a large number of first physiological signals and second physiological signals, further improving the accuracy of the blood pressure prediction results under various prediction conditions, and solving the problem of low blood pressure prediction accuracy under different measurement conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in describing the embodiments of the present application are briefly introduced below.

[0018] Figure 1 A schematic diagram of the system architecture for implementing a blood pressure prediction method provided in an embodiment of the present application; Figure 2 A flow chart of a blood pressure prediction method provided in an embodiment of the present application; Figure 3 A schematic diagram of a flow chart of blood pressure prediction using a prediction model in a blood pressure prediction method provided in an embodiment of the present application; Figure 4 A flow chart of a method for determining target physiological information in a blood pressure prediction method provided in an embodiment of the present application; Figure 5 A schematic diagram of a blood pressure prediction device provided in an embodiment of the present application; Figure 6A schematic diagram of a blood pressure prediction process provided in an embodiment of the present application; Figure 7 A schematic diagram of the structure of a blood pressure prediction device provided in an embodiment of the present application; Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] The embodiments of the present application are described below in conjunction with the drawings in the present application. It should be understood that the implementation methods described below in conjunction with the drawings are exemplary descriptions for explaining the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions of the embodiments of the present application.

[0020] Those skilled in the art will appreciate that, unless expressly stated, the singular forms "one", "an", "" and "the" used herein may also include plural forms. It should be further understood that the terms "including" and "comprising" used in the embodiments of the present application mean that the corresponding features can be implemented as the presented features, information, data, steps, operations, elements and / or components, but do not exclude implementation as other features, information, data, steps, operations, elements, components and / or combinations thereof supported by the technical field. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, the one element may be directly connected or coupled to the other element, or may refer to the one element and the other element establishing a connection relationship through an intermediate element. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein indicates at least one of the items defined by the term, for example, "A and / or B" may be implemented as "A", or as "B", or as "A and B".

[0021] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0022] Here’s the technology: Photoplethysmography sensors are often used in contemporary wearable technology, including fitness bands and smartwatches, to predict blood pressure by detecting changes in light absorption within blood vessels to measure pulse. Despite its widespread use, photoplethysmography’s signals are affected by skin color, sweat, ambient temperature, body hair, and skin thickness, leading to inconsistent readings, especially for those with darker skin or medical conditions. Similarly, wearable pressure sensors that detect pulse wave signals have shown promise in estimating health, but face challenges in accurately aligning invisible arterial lines, limiting their effectiveness in predicting blood pressure for all users.

[0023] In response to at least one of the above-mentioned technical problems or areas that need improvement in the related art, the present application proposes a blood pressure prediction method. The scheme obtains the first physiological information collected by the pulse pressure sensor and the second physiological information collected by the photoplethysmogram sensor within a preset time window, and then selects the physiological information that meets the preset evaluation criteria from the first physiological information and the second physiological information based on multiple evaluation indicators as the target physiological information, inputs the target physiological information into the prediction model, and obtains the systolic pressure and diastolic pressure of the target object output by the prediction model. Before inputting the collected physiological information into the prediction model, based on multiple evaluation indicators, higher-quality physiological information is selected from the first physiological information and the second physiological information as the target physiological information and input into the prediction model, so as to ensure that the physiological information input into the prediction model is accurate and effective, and ensure that high-quality data is used for blood pressure prediction, thereby avoiding the influence of factors such as the user's current physiological characteristics and the environment in which the user is located on the quality of the collected physiological signals, thereby affecting the final prediction result, and the prediction model used to predict blood pressure is trained based on a large number of first physiological signals and second physiological signals, further improving the accuracy of the blood pressure prediction results under various prediction conditions, and solving the problem of low blood pressure prediction accuracy under different measurement conditions.

[0024] The following describes several exemplary embodiments to illustrate the technical solutions of the embodiments of the present application and the technical effects produced by the technical solutions of the present application. It should be noted that the following embodiments can refer to, draw on or combine with each other, and the same terms, similar features and similar implementation steps in different embodiments will not be described repeatedly.

[0025] Figure 1 A schematic diagram of a system architecture for implementing a blood pressure prediction method provided in an embodiment of the present application, wherein the system architecture includes: a terminal 120 and a server 140.

[0026] The terminal 120 installs and runs an application program having a blood pressure prediction method. The terminal 120 is used to input a target physiological signal into a prediction model to obtain a prediction result.

[0027] The terminal 120 is connected to the server 140 via a wireless network or a wired network.

[0028] The server 140 includes at least one of a single server, multiple servers, a cloud computing platform, and a virtualization center. Schematically, the server 140 includes a processor 144 and a memory 142, and the memory 142 includes a display module 1421, a control module 1422, and a receiving module 1423. The server 140 is used to provide background services for the application of the method. Optionally, the server 140 undertakes the main computing work and the terminal 120 undertakes the secondary computing work; or, the server 140 undertakes the secondary computing work and the terminal 120 undertakes the main computing work; or, the server 140 and the terminal 120 adopt a distributed computing architecture for collaborative computing.

[0029] Those skilled in the art will appreciate that the number of the above terminals may be more or less. For example, the above terminal may be only one, or the above terminals may be dozens or hundreds, or more. The embodiment of the present application does not limit the number of terminals and device types.

[0030] The present application provides a method for predicting blood pressure. Figure 2 As shown, the method includes: S101, obtaining first physiological information and second physiological information of a target object within a preset time window.

[0031] In the embodiment of the present application, the first physiological information is information collected by a pulse pressure sensor, and the second physiological information is information collected by a photoplethysmogram sensor.

[0032] In the embodiments of the present application, the pulse pressure sensor is a sensor used to measure the pulse waveform or blood pressure fluctuations in blood vessels, which is commonly used in medical monitoring, especially in the diagnosis and management of cardiovascular diseases. This type of sensor can capture the vascular pressure fluctuations caused by the heartbeat, thereby providing health information such as blood pressure and heart rate. The pulse pressure sensor is mainly used to continuously monitor physiological parameters such as blood pressure, pulse waveform, and vascular elasticity.

[0033] In the embodiment of the present application, the first physiological information is collected by a pulse pressure sensor integrated in the terminal. The first physiological information may be systolic pressure, diastolic pressure, and the difference between systolic pressure and diastolic pressure.

[0034] In the embodiment of the present application, the photoplethysmogram sensor is a sensor commonly used to measure blood oxygen saturation, pulse frequency and blood circulation status. It reflects the pulse fluctuation of the blood by emitting a light beam and detecting the reflection change after the light penetrates the skin, thereby obtaining relevant data on cardiovascular health.

[0035] In the embodiment of the present application, the second physiological information is collected by a photoplethysmogram sensor integrated in the terminal. The second physiological information is also called a PPG (photoplethysmogram) signal, which is a signal that measures the change in blood volume in a blood vessel by optical methods. It infers the blood flow and heart activity by monitoring the change in the intensity of light reflected from the skin surface. The first physiological signal includes the peak value of the pulse, that is, when the heart contracts, the heart pumps blood into the blood vessels, and the speed of blood flow in the arteries increases, resulting in an increase in the blood volume of the blood vessel wall, thereby causing a rising peak value of the PPG signal, the valley value of the pulse, that is, when the heart relaxes, the blood flow slows down, the amount of blood in the blood vessels decreases, and the intensity of the PPG signal decreases, forming a valley value, and the pulse waveform, that is, the waveform of the change in blood volume caused by each heart beat. The cycle of each wave reflects a heartbeat cycle, heart rate, and other information related to cardiovascular characteristics.

[0036] In an embodiment of the present application, the preset time window refers to a continuous sequence of physiological information collected by the sensor within a preset time period that changes with time. For example, if the preset time window is 3 seconds, the first physiological information within the preset time window may include: the difference between the systolic pressure and the diastolic pressure at the first second when the pulse pressure sensor starts collecting, the difference between the systolic pressure and the diastolic pressure at the second second, and the difference between the systolic pressure and the diastolic pressure at the third second; the second physiological information within the preset time window may include: information on the changes in the pulse waveform from the start of collection to the end of collection by the photoelectric volumetric pulse wave sensor.

[0037] In an embodiment of the present application, the target object refers to an object for which blood pressure prediction is to be performed. The target object wears a terminal that integrates a pulse pressure sensor and a photoplethysmogram sensor. Through the sensors on the terminal, the first physiological information and the second physiological information of the target object within a preset time window are obtained.

[0038] S102, based on multiple evaluation indicators, selecting physiological information that meets a preset evaluation standard from the first physiological information and the second physiological information as target physiological information.

[0039] In the embodiment of the present application, in the process of collecting the physiological information of the target object through the pulse pressure sensor and the photoplethysmogram sensor, various factors may affect the quality of the collected physiological information, thereby affecting the accuracy of the prediction result.

[0040] In an embodiment of the present application, the quality of the collected physiological information can be tested from multiple dimensions based on multiple evaluation criteria, so as to ensure that the physiological information used for blood pressure prediction can accurately reflect the user's true physiological state and is valid physiological information, thereby providing a reliable reference for prediction.

[0041] In an embodiment of the present application, the evaluation index for evaluating the quality of physiological information can be the ratio of effective information and noise in the calculated physiological information, the completeness of the physiological information, the stability of the physiological information, the ratio of the peak intensity to the average value of the physiological information, the autocorrelation of the physiological information, the distortion of the physiological information, or any other index that can characterize the quality of the physiological information.

[0042] In the embodiment of the present application, physiological information that meets the preset evaluation criteria means that, based on a comprehensive consideration of multiple evaluation indicators, the quality of physiological information as target physiological information is higher than the quality of another physiological information and can more accurately reflect the true physiological state of the target object.

[0043] In an embodiment of the present application, in order to avoid using physiological information of poor quality for blood pressure prediction, before using the collected physiological information for blood pressure prediction, physiological information that meets preset evaluation criteria is selected from the first physiological information and the second physiological information based on multiple evaluation indicators as the target physiological information for blood pressure prediction.

[0044] In an embodiment of the present application, after the pulse pressure sensor and the photoplethysmogram sensor integrated in the terminal respectively collect the first physiological signal and the second physiological information, the quality and fidelity of the first physiological information and the second physiological information are evaluated in real time, and the highest quality information is selected from the first physiological information and the second physiological information as the target physiological information.

[0045] In an embodiment of the present application, the sensor is connected to an analog-to-digital converter board, which can ensure accurate digitization of analog signals.

[0046] S103, inputting the target physiological information into the prediction model to obtain the prediction result output by the prediction model.

[0047] In the embodiment of the present application, the prediction result includes: the systolic blood pressure and diastolic blood pressure of the target object.

[0048] In an embodiment of the present application, the prediction result can be a single blood pressure prediction value, that is, predicting the instantaneous diastolic and systolic pressures at a specific moment, such as the systolic and diastolic pressures at a certain time point in a preset time window; the prediction result can also be the systolic and diastolic pressures that change with time, such as the systolic and diastolic pressures that change with time within a preset time window, which does not provide a fixed blood pressure value, but a change curve that changes with time.

[0049] In an embodiment of the present application, the prediction model is trained based on multiple training samples and corresponding training labels, the multiple training samples come from a first sample set and a second sample set, and the training labels are the systolic pressure and diastolic pressure corresponding to the training samples; the first sample set is composed of the first physiological information of the sample objects, and the second sample set is composed of the second physiological information of the sample objects.

[0050] In an embodiment of the present application, the first physiological information in the first sample set is collected through a pulse pressure sensor under different skin colors, different exercise conditions and different environmental settings, and the corresponding training labels, namely, systolic pressure and diastolic pressure, are obtained through a traditional cuff method; the second physiological information in the second sample set is also collected through a pulse pressure sensor under different skin colors, different exercise conditions and different environmental settings, and the corresponding training labels, namely, systolic pressure and diastolic pressure, are also obtained through a traditional modification method, and before using the training samples collected under various conditions, the data in the training samples will also be subjected to noise and artifact removal. The training samples and training labels collected through the above scheme can ensure that high-quality training data is provided for the training of the prediction model.

[0051] In an embodiment of the present application, the prediction model is a deep neural network model, which is trained by acquiring physiological information and corresponding blood pressure values ​​under various changing conditions so that the prediction model can recognize complex patterns in the physiological information, that is, recognize changes in physiological state, skin color, movement state and environmental state, and predict the corresponding diastolic and systolic blood pressure.

[0052] In an embodiment of the present application, after the target physiological information is input into the prediction model, the prediction model will predict the blood pressure of the target object based on the target physiological information, thereby obtaining the systolic pressure and diastolic pressure of the target object.

[0053] The blood pressure prediction method provided in the embodiment of the present application combines hardware (pulse pressure sensor and photoplethysmogram sensor) and software algorithm (deep neural network) to provide accurate and continuous beat-by-beat blood pressure prediction, combines multi-sensing technology with real-time physiological information to ensure high-quality data for blood pressure prediction, and utilizes deep learning technology that adapts to various conditions to cope with the impact of information changes and environmental factors, thereby greatly improving the accuracy of blood pressure prediction under different prediction conditions. Compared with traditional cuff-type devices that provide isolated blood pressure measurement values ​​at intervals, the embodiment of the present application provides continuous beat-by-beat blood pressure prediction based on machine learning, and has achieved detection of tiny fluctuations in blood pressure. The continuous data stream helps to detect trends and abnormalities, and can provide a more accurate and detailed assessment of cardiovascular health, thereby capturing abnormalities that cannot be captured by traditional periodic measurements.

[0054] In an embodiment of the present application, a prediction model trained with a large-scale data set is used to process the target physiological information, that is, the prediction model is trained using the physiological information and corresponding blood pressure values ​​in various situations, so that in various physiological conditions and environmental conditions, the blood pressure of the target object can be accurately predicted based on the currently acquired physiological information. Traditional photoplethysmogram sensors usually reduce the accuracy of blood pressure prediction for people with higher melanin levels (people with darker skin colors). The embodiment of the present application ensures that a reliable blood pressure monitoring method is provided for people with different physiological characteristics by combining multiple sensor types, preset evaluation criteria and prediction models trained with large-scale data.

[0055] Based on the above embodiments, as an optional embodiment, the prediction model includes a first sub-model and a second sub-model, the first sub-model is trained based on the training samples in the first sample set and the corresponding training labels; the second sub-model is trained based on the training samples in the second sample set and the corresponding training labels.

[0056] In an embodiment of the present application, since the collected first physiological information and second physiological information are not both used for model prediction, but physiological information with higher quality is selected as target physiological information for prediction, the prediction model includes a first prediction model and a second prediction model. After determining the target physiological information, the corresponding model is selected from the first prediction model and the second prediction model for blood pressure prediction. The first prediction model is used to predict the blood pressure of the target object based on the first physiological information, and the second prediction model is used to predict the blood pressure of the target object based on the second physiological information.

[0057] In an embodiment of the present application, the first sub-model is trained according to the training samples in the first sample set and the corresponding diastolic pressure and systolic pressure, and the second sub-model is trained according to the training samples in the second sample set and the corresponding diastolic pressure and systolic pressure.

[0058] In the embodiment of the present application, if the target physiological information is the first physiological information, the target physiological information is input into the first sub-model; if the target physiological information is the second physiological information, the target physiological information is input into the second sub-model.

[0059] In an embodiment of the present application, after determining the target physiological information, the corresponding prediction model is determined according to the target physiological information. When the target physiological information is the first physiological information, the target physiological information is input into the first sub-model. When the target physiological information is the second physiological information, the target physiological information is input into the second sub-model.

[0060] In the above scheme, in order to ensure that blood pressure prediction can be accurately performed using the first physiological information or the second physiological information, the prediction model is divided into a first sub-model and a second sub-model, and the corresponding sub-model is selected for prediction according to the target physiological information, thereby ensuring the accuracy of blood pressure prediction.

[0061] Based on the above embodiments, as an optional embodiment, the first sub-model includes a first grandchild model and a second grandchild model, and the second sub-model includes a third grandchild model and a fourth grandchild model.

[0062] In the embodiment of the present application, since there is no exact relationship between systolic pressure and diastolic pressure, each type of blood pressure requires a trained model for detection. Therefore, for the prediction of systolic pressure and diastolic pressure, a corresponding model is required for prediction. The first sub-model specifically for blood pressure prediction based on the first physiological information includes a first sub-model and a second sub-model, the first sub-model is used to predict systolic pressure based on the first physiological information, and the second sub-model is used to predict diastolic pressure based on the first physiological information; the second sub-model specifically for blood pressure prediction based on the second physiological information includes a third sub-model and a fourth sub-model, the third sub-model is used to predict systolic pressure based on the second physiological information, and the fourth sub-model is used to predict diastolic pressure based on the second physiological information.

[0063] In the embodiment of the present application, the first submodel is trained based on the training samples in the first sample set and the corresponding first training labels, and the first training labels are the systolic blood pressure of the training samples.

[0064] In the embodiment of the present application, the first grandchild model is trained using training samples in the first sample set and the systolic pressure corresponding to the training samples until the first grandchild model converges.

[0065] In the embodiment of the present application, the second submodel is trained based on the training samples in the first sample set and the corresponding second training labels, and the second training labels are the diastolic pressure of the training samples.

[0066] In the embodiment of the present application, the second grandchild model is trained using the training samples in the first sample set and the diastolic pressure corresponding to the training samples until the second grandchild model converges.

[0067] In the embodiment of the present application, the third submodel is trained based on the training samples in the second sample set and the corresponding third training labels, and the third training labels are the systolic blood pressure of the training samples.

[0068] In the embodiment of the present application, the third grandchild model is trained using the training samples in the second sample set and the systolic pressure corresponding to the training samples until the third grandchild model converges.

[0069] In the embodiment of the present application, the fourth grandson model is trained based on the fourth training label corresponding to the first level of the training sample in the second sample set, and the fourth training label is the diastolic pressure of the training sample.

[0070] In the embodiment of the present application, the fourth grandson model is trained using the training samples in the second sample set and the systolic pressure corresponding to the training samples until the fourth grandson model converges.

[0071] In an embodiment of the present application, the above-mentioned models are trained by minimizing the mean square error using the Adam optimizer to learn, and the Adam optimizer can adjust the learning rate to achieve efficient convergence.

[0072] In the above scheme, since there is no exact relationship between diastolic pressure and systolic pressure, each blood pressure corresponds to a separately trained model, thereby ensuring that whether the first physiological information is used as the target physiological information or the second physiological information is used as the target physiological information, the systolic pressure and diastolic pressure of the target object can be accurately predicted.

[0073] The prediction model provided in the embodiment of the present application is a model integrated by four grandchild models, the model structures of the four grandchild models are completely consistent, and the training methods are also the same (there are differences in the training samples and training labels used). After the target physiological information is input into the prediction model, the corresponding grandchild model is selected to predict the blood pressure according to the information type of the target physiological information and the selected predicted blood pressure. When the target physiological information input into the prediction model is the first physiological information and the systolic pressure is predicted, the prediction model predicts the systolic pressure of the target object based on the first grandchild model and the target physiological information; when the target physiological information input into the prediction model is the first physiological information and the diastolic pressure is predicted, the prediction model predicts the diastolic pressure of the target object based on the second grandchild model and the target physiological information; when the target physiological information input into the prediction model is the second physiological information and the systolic pressure is predicted, the prediction model predicts the systolic pressure of the target object based on the third grandchild model and the target physiological information; when the target physiological information input into the prediction model is the second physiological information and the diastolic pressure is predicted, the prediction model predicts the diastolic pressure of the target object based on the fourth grandchild model and the target physiological information.

[0074] Based on the above embodiments, as an optional embodiment, the prediction model includes: a feature extraction layer, an LSTM (Long Short-Term Memory) layer, and a fully connected layer; refer to Figure 3 As shown, it exemplarily shows a flow chart of blood pressure prediction by the prediction model, and the specific contents are as follows: S201, inputting target physiological information into a feature extraction layer, and obtaining a feature vector sequence of the target physiological information output by the feature extraction layer; S202, inputting the feature vector sequence into the LSTM layer to obtain the time series features of the target physiological information output by the LSTM layer; the time series features are used to reflect the information features of the target physiological information changing over time; S203, input the time series features into the fully connected layer to obtain the prediction result output by the fully connected layer.

[0075] In an embodiment of the present application, the feature extraction layer is composed of three convolutional layers, a normalization layer, a pooling layer and a elimination layer. The three convolutional layers are used to extract features from physiological information; the normalization layer is usually applied after the convolutional layer to standardize the output of each convolutional layer, which helps to alleviate the problems of gradient vanishing and gradient exploding; the role of the pooling layer is to reduce the spatial size of the feature map by downsampling, thereby reducing the amount of calculation and preventing the model from overfitting; the elimination layer is used to reduce overfitting by randomly discarding a part of the neurons during the training process, so that the model does not rely on a specific neuron during the training process, thereby improving the generalization ability of the model.

[0076] In the embodiment of the present application, since the first physiological information and the second physiological information within the preset time window are both continuous time series data, the LSTM layer can effectively capture the time dependency, dynamically adjust the focus on historical information, process noise and nonlinear relationships, thereby making blood pressure prediction more accurate and reliable.

[0077] In an embodiment of the present application, the fully connected layer is used to convert the extracted features into the final prediction results. The fully connected layer maps the input features through linear transformation and activation function, models the relationship between each feature, and thus generates the final prediction results.

[0078] In S201 of the embodiment of the present application, the target physiological information is input into the feature extraction layer, and the feature extraction layer extracts features in the physiological information through multiple convolution operations, thereby obtaining a feature vector sequence of the target physiological information output by the feature extraction layer.

[0079] In S202 of the embodiment of the present application, the feature vector sequence is input into the LSTM layer, and the LSTM layer processes the feature vector based on the forward timing and the backward timing, so as to obtain the timing characteristics of the target physiological information output by the LSTM layer. The timing characteristics can effectively reflect the dynamic changes and timing dependencies in the data.

[0080] In the embodiments of the present application, physiological information exhibits different characteristics at different time points, and the relationship between these characteristics changes dynamically. Therefore, the extraction of time series features can help the model accurately understand the time dependence of physiological information, and thus make more accurate blood pressure predictions.

[0081] In S203 of the embodiment of the present application, the time series features are input into the fully connected layer, the fully connected layer transforms the time series features into a one-dimensional vector, and then the one-dimensional vector identifier of the time series features is mapped to the blood pressure prediction through the fully connected layer, and finally the blood pressure prediction result is obtained.

[0082] In the above scheme, the powerful feature extraction capability of the feature extraction layer and the powerful time series modeling capability of the LSTM layer can be fully utilized to improve the accuracy and generalization of blood pressure prediction, and enhance the model's ability to process complex time series, so that the prediction model can be more accurate when facing dynamically changing physiological information.

[0083] Based on the above embodiments, as an optional embodiment, Figure 4 The figure shows a flow chart of a method for determining target physiological information, and the specific contents are as follows: S301, determining multiple evaluation indicators and the weight of each evaluation indicator; S302, respectively calculating a first index value of each evaluation index of the first physiological information and a second index value of each evaluation index of the second physiological information; S303, determining a first indicator score of the first physiological information based on the first indicator value and the corresponding weight of each evaluation indicator of the first physiological information; S304, determining a second indicator score of the second physiological information based on the second indicator value and the corresponding weight of each evaluation indicator of the second physiological information; S305 , selecting physiological information with a higher index score from the first index score and the second index score as target physiological information.

[0084] In S301 of the embodiment of the present application, the weights of each evaluation indicator can be dynamically set in advance according to the actual application situation of the present application. After obtaining the first physiological information and the second physiological information, the evaluation indicators currently used to evaluate the quality of the physiological information and the weights corresponding to each evaluation indicator are determined.

[0085] In S302 of the embodiment of the present application, for the first physiological information and the second physiological information, it is necessary to determine the first indicator value and the second indicator value of each evaluation indicator respectively. For example, if the evaluation indicator is the signal-to-noise ratio, then it is necessary to determine the first indicator value of the signal-to-noise ratio of the first physiological information and the second indicator value of the signal-to-noise ratio of the second physiological information.

[0086] In S303 of the embodiment of the present application, since a single evaluation indicator is usually not used to compare the quality of physiological information, for each evaluation indicator of the first physiological information, the above evaluation indicator is multiplied by the corresponding weight, and the products of each evaluation indicator and the weight are added to obtain a score characterizing the quality of the physiological information, that is, the first indicator score of the first physiological information is obtained, wherein the sum of the weights of each evaluation indicator is 1.

[0087] In S304 of the embodiment of the present application, for each evaluation indicator of the second physiological information, the above evaluation indicator is multiplied by the corresponding weight, and the products of each evaluation indicator and the weight are added to obtain a score representing the quality of the victory information, that is, the second indicator score of the second physiological information is obtained.

[0088] In the embodiment of the present application, the evaluation index includes at least one of the following: Signal-to-noise ratio of physiological information; Stability of physiological information; Continuity of physiological information.

[0089] In the embodiments of the present application, the signal-to-noise ratio of physiological information refers to a measure of the ratio of effective information intensity to noise intensity. A higher signal-to-noise ratio indicates better information quality, that is, the intensity of effective information is more obvious relative to noise. Conversely, a lower signal-to-noise ratio indicates that the effective information is weaker and the noise is stronger.

[0090] In the embodiments of the present application, the stability of physiological information refers to the variation amplitude and consistency of physiological information over a period of time, and describes the consistency of the fluctuation degree of information over time. The stability of the signal is of great significance for long-term detection and diagnosis.

[0091] In the embodiments of the present application, the stability of physiological information refers to the integrity of the information in the time dimension, which refers to whether the information can be continuously collected without interruption. In long-term tracking and real-time prediction, it helps in the continuous observation and evaluation of physiological processes. Therefore, the continuity of physiological information is crucial.

[0092] In the above scheme, comparisons are made from multiple quality evaluation dimensions based on predefined evaluation indicators (such as signal-to-noise ratio, stability, and continuity) and corresponding weights, and the highest quality physiological information is selected as the target physiological information, which ensures that the most reliable physiological information is used for blood pressure prediction.

[0093] On the basis of the above embodiments, as an optional embodiment, the preset evaluation standard further sets the indicator thresholds of various evaluation indicators. When it is detected that the first indicator value of each evaluation indicator of the first physiological signal and the second indicator value of each evaluation indicator of the second physiological signal do not meet the corresponding indicator thresholds, it is considered that the quality of the signals collected by the pulse pressure sensor and the photoplethysmogram sensor is poor. At this time, an indication information indicating that the quality of the collected information is poor is displayed, instructing the target object to adjust the contact position of the sensor, or to re-collect the first physiological information and the second physiological information.

[0094] In the above scheme, an indicator threshold is set for each evaluation indicator of the first physiological information and the second physiological information to avoid inputting one of the physiological information as the target physiological information into the prediction model for blood pressure prediction when the quality of both the first physiological information and the second physiological information is very poor, thereby ensuring the use of high-quality physiological information for blood pressure prediction and further ensuring the accuracy of the blood pressure prediction results.

[0095] Based on the above embodiments, as an optional embodiment, after the prediction model outputs the prediction result, if the prediction result meets the alarm condition, an alarm message is generated; the alarm information is used to indicate that there is an abnormality in the systolic blood pressure and / or diastolic blood pressure of the target object; the alarm information is displayed in a preset manner.

[0096] In the embodiment of the present application, the systolic pressure and the diastolic pressure each have a corresponding normal range, so the alarm conditions can be set in advance. For example, when the systolic pressure is greater than 120 mmHg, the systolic pressure meets the alarm condition, and when the diastolic pressure is greater than 80 mmHg, the diastolic pressure meets the alarm condition. When it is detected that the prediction result meets the above alarm conditions, an alarm message is displayed to indicate that the corresponding systolic pressure and / or diastolic pressure is abnormal.

[0097] In the embodiment of the present application, the alarm information can be displayed by pushing a message, or by an accompanying vibration effect.

[0098] In the embodiments of the present application, timely feedback of abnormal fluctuations in blood pressure is crucial for users who need to adjust their medications, lifestyles, and health care plans in a timely manner, and can effectively help users take corresponding measures when major changes in blood pressure are detected.

[0099] Based on the above embodiments, the prediction result is sent to the bound terminal; The bound terminal includes at least one of the following: The target subject's medical service provider's terminal; Terminals for the target’s family members; The terminal of the caretaker of the target object.

[0100] In an embodiment of the present application, a bound terminal can be pre-set. After establishing a binding with other terminals, the prediction result can be sent to the bound terminal so that the bound terminal can monitor the blood pressure data of the target object in real time and analyze the blood pressure data.

[0101] In an embodiment of the present application, the bound terminal can be a terminal of a medical service provider of the target object, such as a doctor who provides medical care and other services to the target object, so that medical personnel can monitor the patient's data in real time no matter where they are. The doctor can provide timely feedback and intervention to the patient based on the real-time blood pressure prediction value, thereby improving the patient's treatment effect; the bound terminal can also be a terminal of the target object's family members and caregivers, so that they can share their health indicators with their family members and caregivers immediately.

[0102] refer to Figure 5 As shown, it exemplarily shows a schematic diagram of a blood pressure prediction device, and the specific contents include: In the embodiment of the present application, the blood pressure prediction device includes: a pulse pressure sensor 501, a photoplethysmogram sensor 502, a filter 503, a power regulator 504, an amplifier 505, a battery 506, an analog-to-digital converter 507, a microcontroller 508 and a communication module 509.

[0103] Among them, the pulse pressure sensor 501 is used to collect the first physiological information of the target object; the photoplethysmogram sensor 502 is used to collect the second physiological information of the target object; the filter 503 is used to filter the physiological information collected by the pulse pressure sensor and the photoplethysmogram sensor, remove unnecessary noise and interference, and retain effective physiological information; the power regulator 504 is used to stabilize the voltage and ensure that each component obtains a suitable working voltage. It converts the output voltage of the battery into the voltage required by different components to ensure the stability and reliability of the system; the amplifier 505 is used to amplify the filtered physiological information to improve the information amplitude for digitization and processing; battery 506 is used to output voltage to the power regulator; analog-to-digital converter 507 is used to convert the analog signal of the sensor into a digital signal for processing by the microcontroller unit; microcontroller 508 is used to receive the digital signal from the digital-to-analog converter and perform calculation, control and analysis tasks, such as using the locally stored prediction model to predict blood pressure, or sending the digital signal to the cloud for the cloud to predict blood pressure based on the stored prediction model; communication module 509 is used to wirelessly transmit the physiological information or blood pressure prediction results collected by the device to an external device, or the blood pressure prediction results transmitted by the cloud.

[0104] refer to Figure 6 As shown, it exemplarily shows a flow chart of blood pressure prediction, and the specific contents include: S601, collecting first physiological information and second physiological information of a preset time window through a pulse pressure sensor and a photoplethysmogram sensor respectively; S602, filtering and amplifying the first physiological information and the second physiological information, removing noise in the first physiological information and the second physiological information, and enhancing the information amplitude of the physiological information; S603, selecting physiological information that meets a preset evaluation standard from the first physiological information and the second physiological information as target physiological information according to multiple evaluation indicators; S604, inputting the target physiological information into the prediction model; S605, extracting features of the target physiological information through a feature extraction layer of the prediction model to obtain a feature vector sequence of the target physiological information; S606, processing the feature vector sequence of the target physiological information through the LSTM layer of the prediction model to obtain the time series characteristics of the target physiological information; S607, converting the time series features into final prediction results through the fully connected layer of the prediction model to obtain the systolic pressure and diastolic pressure output by the prediction model; S608, systolic and diastolic blood pressure are displayed.

[0105] In the above scheme, continuous, non-invasive and accurate beat-to-beat blood pressure prediction can be achieved, and dynamic selection is made based on quality between physiological information collected in real time to ensure that the highest quality physiological information is used as input, and it can be adjusted according to user movement, skin color, environmental conditions and other factors that interfere with traditional blood pressure monitoring systems. By integrating deep neural networks, the accuracy, consistency and reliability of cardiovascular health monitoring for different users are further improved. In addition, since there is no need to use expensive cuff-type or clinical-grade equipment, a cost-effective alternative is provided for personal and clinical use. While ensuring the accuracy of blood pressure prediction, the cost of use is reduced, so that users can afford continuous blood pressure testing without sacrificing the accuracy and reliability of blood pressure prediction, and the design of this application is convenient for flexible positioning on the body. This not only improves the user's comfort, but also improves the consistency of physiological information collection. The system can adapt to subtle changes in the position of the sensor, so it is very suitable for daily use, and users do not need to master a lot of technical knowledge.

[0106] The present application embodiment provides a blood pressure prediction device, such as Figure 7 As shown, the device prediction apparatus 70 may include: an acquisition module 701 , a selection module 702 and an input module 703 .

[0107] Specifically, the acquisition module 701 is used to acquire the first physiological information and the second physiological information of the target object within a preset time window; the first physiological information is the information collected by the pulse pressure sensor, and the second physiological information is the information collected by the photoplethysmogram sensor; A selection module 702, configured to select physiological information that meets a preset evaluation standard from the first physiological information and the second physiological information as target physiological information based on multiple evaluation indicators; The input module 703 is used to input the target physiological information into the prediction model to obtain the prediction results output by the prediction model; the prediction results include: the systolic pressure and diastolic pressure of the target object; The prediction model is trained based on multiple training samples and corresponding training labels, the multiple training samples come from the first sample set and the second sample set, and the training labels are systolic pressure and diastolic pressure corresponding to the training samples; The first sample set is composed of first physiological information of the sample subject, and the second sample set is composed of second physiological information of the sample subject.

[0108] The device of the embodiments of the present application can execute the method provided by the embodiments of the present application, and the implementation principles are similar. The actions performed by each module in the device of each embodiment of the present application correspond to the steps in the method of each embodiment of the present application. For the detailed functional description of each module of the device, please refer to the description in the corresponding method shown in the previous text, which will not be repeated here.

[0109] Further, in a possible implementation, the prediction model includes a first sub-model and a second sub-model, and the first sub-model is trained according to the training samples in the first sample set and the corresponding training labels; The second sub-model is trained based on the training samples in the second sample set and the corresponding training labels; If the target physiological information is the first physiological information, the target physiological information is input into the first sub-model; If the target physiological information is the second physiological information, the target physiological information is input into the second sub-model.

[0110] In another possible implementation, the first sub-model includes a first grandchild model and a second grandchild model, and the second sub-model includes a third grandchild model and a fourth grandchild model; The first submodel is trained based on the training samples in the first sample set and the corresponding first training labels, and the first training labels are the systolic blood pressure of the training samples; The second Sun model is trained based on the training samples in the first sample set and the corresponding second training labels, and the second training labels are the diastolic pressure of the training samples; The third Sun model is trained based on the training samples in the second sample set and the corresponding third training labels, where the third training labels are the systolic blood pressure of the training samples; The fourth grandson model is trained according to the fourth training label corresponding to the first level of the training sample in the second sample set, and the fourth training label is the diastolic pressure of the training sample.

[0111] In another possible implementation, the prediction model includes: a feature extraction layer, a long short-term memory (LSTM) layer, and a fully connected layer; inputting the target physiological information into the feature extraction layer to obtain a feature vector sequence of the target physiological information output by the feature extraction layer; The feature vector sequence is input into the LSTM layer to obtain the time series characteristics of the target physiological information output by the LSTM layer; the time series characteristics are used to reflect the information characteristics of the target physiological information changing over time; Input the time series features into the fully connected layer to obtain the prediction results output by the fully connected layer.

[0112] In yet another possible implementation, a plurality of evaluation indicators and a weight of each evaluation indicator are determined; respectively calculating a first index value of each evaluation index of the first physiological information and a second index value of each evaluation index of the second physiological information; determining a first indicator score of the first physiological information based on first indicator values ​​and corresponding weights of each evaluation indicator of the first physiological information; determining a second indicator score of the second physiological information based on the second indicator value and the corresponding weight of each evaluation indicator of the second physiological information; From the first indicator score and the second indicator score, select the physiological information with a higher indicator score as the target physiological information; The evaluation indicators include at least one of the following: Signal-to-noise ratio of physiological information; Stability of physiological information; Continuity of physiological information.

[0113] In another possible implementation, after obtaining the prediction result output by the prediction model, if the prediction result meets the alarm condition, an alarm message is generated; the alarm message is used to indicate that the systolic blood pressure and / or diastolic blood pressure of the target object is abnormal; Display warning information in a preset manner.

[0114] In yet another possible implementation, the prediction result is sent to the bound terminal; The bound terminal includes at least one of the following: The target subject's medical service provider's terminal; Terminals for the target’s family members; The terminal of the caretaker of the target object.

[0115] In an embodiment of the present application, an electronic device (computer device / equipment / system) is provided, including a memory, a processor, and a computer program stored in the memory. The processor executes the above-mentioned computer program to implement the steps of the blood pressure prediction method. Compared with the related art, it can achieve: by acquiring the first physiological information collected by the pulse pressure sensor within a preset time window and the second physiological information collected by the photoplethysmogram sensor, and then based on multiple evaluation indicators, selecting the physiological information that meets the preset evaluation criteria from the first physiological information and the second physiological information as the target physiological information, inputting the target physiological information into the prediction model, obtaining the systolic and diastolic blood pressure of the target object output by the prediction model, and then inputting the collected physiological information into the prediction model. Before inputting into the prediction model, based on multiple evaluation indicators, higher quality physiological information is selected from the first physiological information and the second physiological information as the target physiological information to be input into the prediction model, so as to ensure that the physiological information input into the prediction model is accurate and effective, and to ensure that high-quality data is used for blood pressure prediction, thereby avoiding the quality of the collected physiological signals being affected by factors such as the user's current physiological characteristics and the environment, thereby affecting the final prediction result. The prediction model for predicting blood pressure is trained based on a large number of first physiological signals and second physiological signals, further improving the accuracy of the blood pressure prediction results under various prediction conditions, and solving the problem of low blood pressure prediction accuracy under different measurement conditions. In an optional embodiment, an electronic device is provided, such as Figure 8 As shown, Figure 8 The electronic device 4000 shown includes: a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, such as through a bus 4002. Optionally, the electronic device 4000 may also include a transceiver 4004, which may be used for data interaction between the electronic device and other electronic devices, such as data transmission and / or data reception. It should be noted that in actual applications, the transceiver 4004 is not limited to one, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present application.

[0116] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. Processor 4001 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0117] The bus 4002 may include a path to transmit information between the above components. The bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 4002 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0118] The memory 4003 may be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage media, other magnetic storage devices, or any other medium that can be used to carry or store computer programs and can be read by a computer, without limitation herein.

[0119] The memory 4003 is used to store the computer program for executing the embodiment of the present application, and the execution is controlled by the processor 4001. The processor 4001 is used to execute the computer program stored in the memory 4003 to implement the steps shown in the above method embodiment.

[0120] Among them, the electronic equipment package may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 8 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0121] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps and corresponding contents of the aforementioned method embodiment can be implemented. Compared with the prior art, the present invention can achieve the following: by acquiring the first physiological information collected by the pulse pressure sensor and the second physiological information collected by the photoplethysmogram sensor within a preset time window, and then based on multiple evaluation indicators, selecting the physiological information that meets the preset evaluation criteria from the first physiological information and the second physiological information as the target physiological information, inputting the target physiological information into the prediction model, and obtaining the systolic pressure and diastolic pressure of the target object output by the prediction model; before inputting the collected physiological information into the prediction model, firstly selecting the physiological information with higher quality from the first physiological information and the second physiological information as the target physiological information based on multiple evaluation indicators and inputting it into the prediction model, so as to ensure that the physiological information input into the prediction model is accurate and effective, and ensure that high-quality data is used for the prediction of blood pressure, thereby avoiding the influence of factors such as the current physiological characteristics of the user and the environment in which the user is located on the quality of the collected physiological signals, thereby affecting the final prediction result; and the prediction model used to predict blood pressure is trained based on a large number of first physiological signals and second physiological signals, further improving the accuracy of the blood pressure prediction results under various prediction conditions, and solving the problem of low blood pressure prediction accuracy under different measurement conditions.

[0122] It should be noted that the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0123] The present application also provides a computer program product, including a computer program, which can implement the steps and corresponding contents of the above method embodiments when executed by a processor. Compared with the prior art, it can achieve: By acquiring the first physiological information collected by the pulse pressure sensor and the second physiological information collected by the photoelectric volumetric pulse wave sensor within a preset time window, and then based on multiple evaluation indicators, selecting the physiological information that meets the preset evaluation criteria from the first physiological information and the second physiological information as the target physiological information, the target physiological information is input into the prediction model to obtain the systolic and diastolic pressures of the target object output by the prediction model. Before the collected physiological information is input into the prediction model, based on multiple evaluation indicators, higher-quality physiological information is selected from the first physiological information and the second physiological information as the target physiological information to be input into the prediction model, to ensure that the physiological information input into the prediction model is accurate and effective, and to ensure that high-quality data is used for blood pressure prediction, thereby avoiding the quality of the collected physiological signals affected by factors such as the user's current physiological characteristics and the environment in which they are located, thereby affecting the final prediction result. The prediction model used to predict blood pressure is trained based on a large number of first physiological signals and second physiological signals, further improving the accuracy of the blood pressure prediction results under various prediction conditions, and solving the problem of low blood pressure prediction accuracy under different measurement conditions.

[0124] The terms "first", "second", "third", "fourth", "1", "2", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than that shown or described in the drawings.

[0125] It should be understood that, although each operation step is indicated by arrows in the flowchart of the embodiment of the present application, the implementation order of these steps is not limited to the order indicated by the arrows. Unless clearly stated herein, in some implementation scenarios of the embodiment of the present application, the implementation steps in each flowchart can be performed in other orders according to demand. In addition, some or all of the steps in each flowchart may include multiple sub-steps or multiple stages based on actual implementation scenarios. Some or all of these sub-steps or stages may be executed at the same time, and each sub-step or stage in these sub-steps or stages may also be executed at different times respectively. In different scenarios of execution time, the execution order of these sub-steps or stages may be flexibly configured according to demand, and the embodiment of the present application does not limit this.

[0126] The above are only optional implementation methods for some implementation scenarios of the present application. It should be pointed out that for ordinary technicians in this technical field, without departing from the technical concept of the scheme of the present application, other similar implementation methods based on the technical ideas of the present application are also within the protection scope of the embodiments of the present application.

Claims

1. A blood pressure prediction method, characterized in that: include: Acquiring first physiological information and second physiological information of a target object within a preset time window; The first physiological information is information collected by a pulse pressure sensor, and the second physiological information is information collected by a photoplethysmogram sensor; Based on multiple evaluation indicators, selecting physiological information that meets a preset evaluation standard from the first physiological information and the second physiological information as target physiological information; Inputting the target physiological information into a prediction model to obtain a prediction result output by the prediction model; the prediction result includes: the systolic pressure and diastolic pressure of the target object; The prediction model is trained based on a plurality of training samples and corresponding training labels, the plurality of training samples are from a first sample set and a second sample set, and the training labels are systolic pressure and diastolic pressure corresponding to the training samples; The first sample set is composed of first physiological information of the sample subject, and the second sample set is composed of second physiological information of the sample subject.

2. The method according to claim 1, characterized in that The step of selecting physiological information that meets a preset evaluation standard from the first physiological information and the second physiological information as target physiological information based on the multiple evaluation indicators includes: Determining the multiple evaluation indicators and the weight of each evaluation indicator; respectively calculating a first index value of each evaluation index of the first physiological information and a second index value of each evaluation index of the second physiological information; determining a first indicator score of the first physiological information based on first indicator values ​​and corresponding weights of each evaluation indicator of the first physiological information; determining a second indicator score of the second physiological information based on the second indicator value and the corresponding weight of each evaluation indicator of the second physiological information; From the first index score and the second index score, select physiological information with a higher index score as target physiological information; The evaluation index includes at least one of the following: a signal-to-noise ratio of the physiological information; stability of the physiological information; The continuity of the physiological information.

3. The method according to claim 1, characterized in that: The prediction model includes a first sub-model and a second sub-model, wherein the first sub-model is trained according to the training samples in the first sample set and the corresponding training labels; The second sub-model is trained according to the training samples in the second sample set and the corresponding training labels; The step of inputting the target physiological information into a prediction model comprises: If the target physiological information is the first physiological information, inputting the target physiological information into the first sub-model; If the target physiological information is the second physiological information, the target physiological information is input into the second sub-model.

4. The method according to claim 1, characterized in that The prediction model includes: a feature extraction layer, a long short-term memory (LSTM) layer, and a fully connected layer; The step of inputting the target physiological information into a prediction model to obtain a prediction result output by the prediction model includes: Inputting the target physiological information into the feature extraction layer to obtain a feature vector sequence of the target physiological information output by the feature extraction layer; Inputting the feature vector sequence into the LSTM layer to obtain the time series characteristics of the target physiological information output by the LSTM layer; the time series characteristics are used to reflect the information characteristics of the target physiological information changing over time; The time series features are input into the fully connected layer to obtain the prediction result output by the fully connected layer.

5. The method according to claim 1, characterized in that Obtaining the prediction result output by the prediction model, and then further comprising: If the prediction result meets the alarm condition, an alarm message is generated; the alarm message is used to indicate that the systolic blood pressure and / or diastolic blood pressure of the target object is abnormal; The warning information is displayed in a preset manner.

6. The method according to claim 1, characterized in that The method comprises: Sending the prediction result to the bound terminal; The bound terminal includes at least one of the following: A terminal of a medical service provider of the target object; Terminals of the target object's family members; A terminal of a caregiver of the target object.

7. A blood pressure prediction device, characterized in that: include: An acquisition module, used to acquire first physiological information and second physiological information of a target object within a preset time window; The first physiological information is information collected by a pulse pressure sensor, and the second physiological information is information collected by a photoplethysmogram sensor; A selection module, configured to select, based on a plurality of evaluation indicators, physiological information that meets a preset evaluation standard from the first physiological information and the second physiological information as target physiological information; An input module, used to input the target physiological information into a prediction model to obtain a prediction result output by the prediction model; the prediction result includes: the systolic pressure and diastolic pressure of the target object; The prediction model is trained based on a plurality of training samples and corresponding training labels, the plurality of training samples are from a first sample set and a second sample set, and the training labels are systolic pressure and diastolic pressure corresponding to the training samples; The first sample set is composed of first physiological information of the sample subject, and the second sample set is composed of second physiological information of the sample subject.

8. An electronic device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images