WO2018129822A1 - 一种血压测量方法及设备 - Google Patents
一种血压测量方法及设备 Download PDFInfo
- Publication number
- WO2018129822A1 WO2018129822A1 PCT/CN2017/080811 CN2017080811W WO2018129822A1 WO 2018129822 A1 WO2018129822 A1 WO 2018129822A1 CN 2017080811 W CN2017080811 W CN 2017080811W WO 2018129822 A1 WO2018129822 A1 WO 2018129822A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- blood pressure
- training data
- type
- prediction model
- pressure measurement
- Prior art date
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/021—Measuring pressure in heart or blood vessels
Definitions
- the present application relates to the field of terminals, and in particular, to a blood pressure measuring method and device.
- Blood pressure is the driving force that circulates blood in the blood vessels, providing enough blood for each tissue to maintain normal metabolism.
- high blood pressure showed high blood pressure, is a very common cardiovascular disease, there are about 160 million high blood pressure patients in China, high blood pressure will bring many hazards such as stroke, blindness and myocardial infarction.
- the wearable device measures the blood pressure based on the acquired biological signal, and the working principle is:
- the existing wearable blood pressure device preprocesses the biosignal to obtain the corresponding feature data when measuring the blood pressure, and then directly inputs the feature data into a single prediction model for blood pressure prediction, but a single The predictive model is not suitable for the diversity of blood pressure, and the need to accurately predict blood pressure is not achieved, resulting in low accuracy of blood pressure measurement.
- the embodiment of the present application provides a blood pressure measuring method and device for solving the problem that the blood pressure measurement accuracy is not high.
- the present application provides a blood pressure measurement method, including: a measurement device extracts feature data in a biosignal according to a collected biosignal of a user, and calculates a classification reference value according to the feature data and a preset classification parameter, and then according to the classification
- the preset threshold range into which the reference value falls, the target blood pressure prediction model is determined from the blood pressure prediction model set, and finally the feature data is input into the target blood pressure prediction model to obtain the user's blood pressure measurement result.
- Each of the blood pressure preset models corresponds to a preset threshold range, and the preset threshold range corresponding to the target blood pressure prediction model is the same as the preset threshold range in which the classification reference value falls.
- the measuring device collects the user's biological signal, extracts the feature data from the biological signal, and calculates the classification reference value according to the preset classification parameter and the feature data extracted from the biological signal.
- the preset threshold range in which the classification reference value falls is different
- the target blood pressure prediction model determined by the measuring device is also different. Therefore, the blood pressure measurement method provided by the present application can better adapt to the scene of blood pressure diversity, and the feature is not directly used.
- the data is input to a single blood pressure prediction model, which can effectively improve the accuracy of blood pressure measurement.
- the preset classification parameter is obtained by the following method: the measurement device acquires the training data set, and divides the training data set into at least two according to the actual blood pressure measurement result included in each training data.
- a type of training data subset obtains preset classification parameters according to at least two first type training data subsets and a preset classification function.
- the preset classification function may be a logistic regression function.
- each training data in the training data set includes a characteristic data and an actual blood pressure measurement result
- the actual blood pressure measurement result includes systolic blood pressure and diastolic blood pressure
- the biosignal corresponding to the characteristic data in each training data and the actual blood pressure measurement result The corresponding biosignals are the same.
- the training data set includes a large amount of training data
- the feature data in the kth training data is the feature data extracted for the kth biological signal
- the actual blood pressure measurement result in the kth training data is for the first
- the blood pressure measurement result of k biosignals may refer to a blood pressure measurement result measured by a user using a tool such as a mercury sphygmomanometer or a cuff sphygmomanometer
- k is a positive integer.
- the preset threshold range includes two or more, two or more preset classification parameters are needed, and the two or more classification reference values that need to be calculated are combined to respectively correspond to different preset threshold ranges.
- the blood pressure prediction model set is obtained by the following method:
- the measuring device establishes a blood pressure prediction model according to each of the first type of training data subsets, and obtains a blood pressure prediction model set. Therefore, the measuring device can directly establish a blood pressure prediction model according to the first type of training data subset, and the method is simple.
- the blood pressure prediction model set is obtained by the following method:
- the measuring device divides the training data set into at least two second types of training data subsets according to actual blood pressure measurement results included in each training data, the number of the first type of training data subsets and the second type of training data subsets The number is the same; a blood pressure prediction model is established according to each of the second type of training data subsets, and a blood pressure prediction model set is obtained. Therefore, the measuring device can establish a blood pressure prediction model by using a second subset of training data subsets different from the first type of training data subset, and the method is flexible to suit the needs of different scenarios.
- the number of the first type of training data subset is the same as the number of the second type training data subset, and each of the first type training data subsets corresponds to a first type of actual blood pressure measurement result.
- each second type of training data subset corresponds to a second type of actual blood pressure measurement result range
- the i-th first type of training data subset corresponding to the i-th first type of actual blood pressure measurement result range is A subset of the range of values of the i-th second type of actual blood pressure measurement corresponding to the i-th second-type training data subset, wherein i is a positive integer.
- the reason for dividing the training data set into at least two second types of training data subsets and not continuing to use at least two first type training data subsets is that the training data subsets used to establish different blood pressure prediction models are There is a certain degree of redundancy between them to ensure the accuracy of blood pressure measurements near the classification nodes.
- the present application provides a blood pressure measuring device, including: a collector, a memory, and a processor, wherein the collector is configured to collect a biosignal of the user; the memory is configured to store the computer program; and the processor is configured to invoke The computer program stored in the memory is executed to: extract feature data in the biosignal according to the biosignal of the user collected by the collector; calculate the classification reference value according to the feature data and the preset classification parameter; and preset according to the classification reference value a threshold range, wherein the target blood pressure prediction model is determined from the blood pressure prediction model set, each blood pressure preset model corresponding to a preset threshold range, and the preset threshold range corresponding to the target blood pressure prediction model is the same as the preset threshold range in which the classification reference value falls The characteristic data is input into the target blood pressure prediction model to obtain the blood pressure measurement result of the user.
- the processor is further configured to: acquire a training data set, and train the data set Each piece of training data includes one feature data and one actual blood pressure measurement result, and the biosignal corresponding to the feature data in each training data is the same as the biosignal corresponding to the actual blood pressure measurement result; according to the actual blood pressure measurement result included in each training data And dividing the training data set into at least two first types of training data subsets; and obtaining preset classification parameters according to at least two first type training data subsets and a preset classification function.
- the processor is further configured to: establish a blood pressure prediction model according to each of the first types of training data subsets, and obtain a blood pressure prediction model set.
- the processor is further configured to: divide the training data set into at least two second types of training data subsets according to actual blood pressure measurement results included in each training data, and the first type of training The number of data subsets is the same as the number of the second type of training data subsets; a blood pressure prediction model is established according to each second type of training data subset to obtain a blood pressure prediction model set.
- the number of the first type of training data subset is the same as the number of the second type training data subset, and each of the first type training data subsets corresponds to a first type of actual blood pressure measurement result.
- each second type of training data subset corresponds to a second type of actual blood pressure measurement result range
- the i-th first type of training data subset corresponding to the i-th first type of actual blood pressure measurement result range is A subset of the range of values of the i-th second type of actual blood pressure measurement corresponding to the i-th second-type training data subset, wherein i is a positive integer.
- the present application provides a computer readable storage medium having stored therein instructions that, when executed on a computer, cause the computer to perform the methods described in the various aspects above.
- the present application provides a computer program product comprising instructions for causing a computer to perform the methods described in the above aspects when the computer program product runs on a computer.
- FIG. 1 is a schematic view of a wristwatch having a blood pressure measuring function in an embodiment of the present application
- FIG. 2 is a schematic diagram of a basic structure of a measuring device in an embodiment of the present application.
- FIG. 3 is a flow chart showing an overview of a blood pressure measuring method in an embodiment of the present application.
- FIG. 4 is a specific flowchart of a blood pressure measurement performed by a measuring device according to an embodiment of the present application
- FIG. 5 is a schematic diagram of data distribution of a training data set divided into two second types of training data subsets according to an embodiment of the present application
- FIG. 6 is a schematic diagram of obtaining a corresponding blood pressure prediction model according to a second type of training data subset division result according to an embodiment of the present application.
- the measuring device mentioned in the embodiment of the present application generally refers to a wearable device.
- the wearable device may be a watch, a wristband, etc., as shown in FIG. 1 , a watch with a blood pressure measuring function. .
- FIG. 2 is a schematic diagram showing the basic configuration of a measuring device, which includes: a collector 210, a memory 220, a processor 230, a power source 240, and the like.
- the measurement device may further include: a display unit 250, a Bluetooth module 260, and the like.
- the measuring device mentioned in the embodiment of the present application is basically the same as the measuring device shown in FIG. 2 above. It should be noted that the embodiment of the present application does not relate to the improvement of the hardware configuration of the measuring device.
- the embodiment of the present application provides a blood pressure measurement method to solve the problem that the blood pressure measurement accuracy is not high, and the method includes:
- Step 300 The measuring device extracts feature data in the biosignal according to the collected biometric signal of the user.
- the biosignal here may be an electrocardiographic signal, or a pulse wave signal or the like.
- the feature data extracted from the biosignal may be characteristic data capable of characterizing the biosignal such as the interval between the peak troughs and the full width at half maximum of the crest.
- the general method of extracting the feature data in the biosignal is to first extract the position of the feature point in the biosignal, such as the peak, trough and maximum slope point of the pulse wave signal, and then obtain the feature data based on the position of the feature point and the biosignal.
- the feature data extracted from the biosignal can be written as x0, x1, x2, ..., xn.
- Step 310 The measuring device calculates the classification reference value according to the feature data and the preset classification parameter.
- the preset classification parameter here can be A, which is denoted as a0, a1, a2, ..., an.
- A which is denoted as a0, a1, a2, ..., an.
- the classification reference value calculated by the measuring device is a0*x0+a1*x1+a2*x2+...+an*xn.
- Step 320 The measuring device determines the target blood pressure prediction model from the blood pressure prediction model set according to the preset threshold range in which the classification reference value falls.
- Each blood pressure preset model corresponds to a preset threshold range, and the preset threshold range corresponding to the target blood pressure prediction model is the same as the preset threshold range in which the classification reference value falls.
- the blood pressure prediction model set also includes two, respectively, the corresponding classification reference value is greater than or equal to 0 and the classification reference value is less than 0. .
- the measuring device uses the blood pressure prediction model whose classification reference value is greater than or equal to 0 as the target blood pressure prediction model.
- the classification reference value calculated in step 310 is less than 0, the measuring device will The blood pressure prediction model whose classification reference value is smaller than 0 is used as the target blood pressure prediction model.
- the preset threshold range includes two or more, two or more preset classification parameters are needed, and the two or more classification reference values that need to be calculated are combined to respectively correspond to different preset threshold ranges.
- Step 330 The measuring device inputs the feature data into the target blood pressure prediction model to obtain the blood pressure measurement result of the user.
- the measuring device determines the target blood pressure prediction model, that is, the blood pressure preset model applicable to the current feature data, and then inputs the feature data to the target blood pressure prediction model.
- each blood pressure prediction model in the blood pressure prediction model set has the characteristic data as an independent variable and the blood pressure value as a dependent variable.
- the measuring device collects the biosignal of the user, extracts the feature data from the biosignal, calculates the classification reference value according to the preset classification parameter and the feature data extracted from the biosignal, and determines according to the classification reference value.
- the target blood pressure prediction model inputs the characteristic data to the target blood pressure prediction model to obtain a blood pressure measurement result.
- the preset threshold range in which the classification reference value falls is different, the target blood pressure prediction model determined by the measuring device is also different. Therefore, the blood pressure measurement method provided by the embodiment of the present application can better adapt to the scene of blood pressure diversity, and is no longer directly
- the trait data is input into a single blood pressure prediction model, which effectively improves the accuracy of blood pressure measurement.
- the preset classification parameters mentioned in step 310 and the blood pressure prediction model set mentioned in step 320 are obtained before the measurement device performs blood pressure measurement, and are saved in the measurement device.
- the preset classification parameter and the blood pressure prediction model set may be obtained by establishing a classification model and a blood pressure prediction model, and may be, but not limited to, the following possible implementation manners.
- the measurement device acquires a training data set.
- each training data in the training data set includes a characteristic data and an actual blood pressure measurement result
- the actual blood pressure measurement result includes systolic blood pressure and diastolic blood pressure
- the biosignal corresponding to the characteristic data in each training data The biosignal corresponding to the blood pressure measurement is the same.
- the training data set includes a large amount of training data
- the feature data in the kth training data is the feature data extracted for the kth biological signal
- the actual blood pressure measurement result in the kth training data is for the first
- the blood pressure measurement result of k biosignals, the actual blood pressure measurement result herein may refer to a blood pressure measurement result measured by a user using a tool such as a mercury sphygmomanometer or a cuff sphygmomanometer
- k is a positive integer.
- the measuring device establishes a classification model according to the training data set, and obtains a preset classification parameter. Specifically, the measuring device divides the training data set into at least two first types of training data subsets according to the actual blood pressure measurement results included in each training data, according to at least two first types of training data subsets and preset classifications. Function to get the preset classification parameters. Wherein, each of the first type of training data subsets corresponds to a range of values of the first type of actual blood pressure measurement results.
- the systolic blood pressure (SBP) is equal to 140
- the training data set is divided into two first training data subsets
- the first first training data subset includes the training data with an SBP less than 140.
- the training data, the second first type of training data subset includes training data with an SBP greater than or equal to 140 in the training data.
- the training data with the SBP less than 140 in the training data is marked as -1
- the training data with the SBP greater than or equal to 140 in the training data is marked as +1
- the preset classification parameter A is obtained by the preset classification function, and is recorded as a0, a1. , a2,...,an.
- the above training data set is divided into two first types of training data subsets with SBP equal to 140, wherein the first first type training data subset represents a non-high voltage subset, and the second The first type of training data subset represents a high-voltage subset, which is only an example, and other classifications may be performed according to actual conditions.
- the SBP is equal to 100
- the training data set is divided into two first training data subsets, and the first first training data subset includes training data with SBP less than 100 in the training data, indicating a low voltage subset.
- the second first type of training data subset includes training data in which the SBP is greater than or equal to 100 in the training data, indicating a non-low voltage subset.
- the training data set can also be divided into any two subsets of the first type of training data by using SBP equal to 110, 120, etc., regardless of whether there is a specific meaning such as high voltage and low voltage.
- the measuring device establishes a blood pressure prediction model according to each of the first type of training data subsets, and obtains a blood pressure prediction model set.
- the measuring device divides the training data set into two first training data subsets with the SBP equal to 140, and the preset classification parameters have been obtained according to the two first training data subsets and the preset classification function. Further, the measuring device obtains two blood pressure prediction models according to the two first training data subsets by using a preset regression function, which are respectively recorded as Model 1 and Model 2, for example, the regression function may be a linear regression function, wherein the model The parameter B of 1, denoted as b0, b1, b2, ..., bn, the classification reference value corresponding to model 1 is less than 0, the parameter C of model 2, denoted as c0, c1, c2, ..., cn, the classification corresponding to model 2 The reference value is greater than or equal to zero.
- the measuring device uses the model 1 as the target blood pressure prediction model, and obtains the blood pressure measurement result of the user as b0* X0+b1*x1+b2*x2+...+bn*xn; when the classification reference value calculated by the measuring device is greater than or equal to 0, the measuring device uses model 2 as the target blood pressure prediction model, and obtains the user's blood pressure measurement result as c0*x0. +c1*x1+c2*x2+...+cn*xn.
- the embodiment is exemplified by systolic blood pressure, and the manner of predicting diastolic blood pressure is similar to the method of predicting systolic blood pressure, except that the dependent variable when the model is established is changed from systolic blood pressure to diastolic blood pressure.
- obtaining the blood pressure prediction model set may also adopt the following scheme:
- the measuring device divides the training data set into at least two second types of training data subsets according to actual blood pressure measurement results included in each training data, and establishes a blood pressure prediction model according to each second type training data subset to obtain blood.
- a set of pressure prediction models wherein, each second type of training data subset corresponds to a second type of actual blood pressure measurement result value range.
- the number of the first type of training data subset is the same as the number of the second type training data subset, and the i-th first type training data subset corresponds to the ith first type of actual blood pressure measurement result range.
- the measuring device divides the training data set into two second types of training data subsets according to actual blood pressure measurement results included in each training data, as shown in FIG. 5, the first second type training data subset includes The training data has an SBP less than 150 training data, and the second second type training data subset includes training data with an SBP greater than or equal to 130 in the training data, and then the measuring device presets according to the two second types of training data subsets.
- the regression function obtains two blood pressure prediction models, which are model x and model y, respectively, as shown in Fig. 6.
- model x corresponds to the first first type of training data subset
- model x parameter B is recorded as b0, b1.
- model x also corresponds to the classification reference value is less than 0
- model y corresponds to the second first type of training data subset
- model y parameter C denoted as c0, c1, c2, ..., cn
- model y also corresponds to the classification reference value is greater than or equal to 0.
- the training data set is divided into at least two second types of training data subsets, and the reason for not using the at least two first type training data subsets is that the different blood pressure prediction models are used. There is a certain degree of redundancy between the training data subsets to ensure the accuracy of blood pressure measurements near the classification nodes.
- the training data set is divided into four first training data subsets by dividing SBP equal to 100 and SBP equal to 140.
- the first first training data subset includes training data with SBP less than 100 in the training data.
- the second first type training data subset includes training data in which the SBP is greater than or equal to 100 in the training data
- the third first type training data subset includes training data in which the SBP is less than 140 in the training data
- the fourth first type training The data subset includes training data with an SBP greater than or equal to 140 in the training data.
- the preset classification parameter A1 is obtained through a preset classification function.
- the preset classification parameter A2 is obtained by a preset classification function.
- the training data set is divided into three second types of training data subsets, and the first second type training data subset includes training data with an SBP less than 100 in the training data, and a second first training data subset.
- the training data includes training data with an SBP greater than or equal to 100 and an SBP less than 140
- the third first training data subset includes training data with an SBP greater than or equal to 140 in the training data
- the subset of the training data based on the three second types is passed.
- the preset regression function obtains three blood pressure prediction models, namely model x1, model x2, and model x3.
- a certain degree of redundancy can also be set between the respective training data subsets.
- the model x1 corresponds to the classification reference value calculated according to A1 is less than 0, the model x2 corresponding to the classification reference value calculated according to A1 is greater than or equal to 0, and the classification reference value is less than 0 according to A2 calculation, and the model x3 correspondingly obtains the classification reference value according to A2. Equal to 0.
- the measuring device calculates the classification reference value according to A1, denoted as B1, and calculates the classification reference value according to A2, which is recorded as B2.
- B1 is less than 0, the measuring device takes x1 as The target blood pressure prediction model, when B1 is greater than or equal to 0 and B2 is less than 0, the measuring device uses x2 as the target blood pressure prediction model, and when B3 is greater than or equal to 0, the measuring device uses x3 as the target blood pressure prediction model.
- the foregoing process of obtaining the preset classification parameter and the blood pressure prediction model set may be performed by the measurement device or by other devices. After the other device performs the above process to obtain the preset classification parameter and the blood pressure prediction model set, The preset classification parameters and the blood pressure prediction model set are stored in the measurement device.
- the measuring device calculates the classification reference value according to the feature data and the preset classification parameter, and the measuring device And determining, according to the preset threshold range in which the classification reference value falls, at least two target blood pressure prediction models, wherein the probability sum of the at least two target blood pressure prediction models is 1, and the measuring device inputs the feature data To each target blood pressure prediction model, the at least two target blood pressure prediction models are respectively obtained corresponding to blood pressure measurement results, and then the final user's blood pressure measurement results are determined according to the probability of each target blood pressure prediction model.
- the measuring device determines two target blood pressure prediction models from the blood pressure prediction model set according to the preset threshold range into which the classification reference value falls, respectively, the model 1 and the model 2, the probability of the model 1 is 70%, and the probability of the model 2 At 30%, the measuring device inputs the characteristic data to the two target blood pressure prediction models respectively, obtains two blood pressure measurement results, respectively S1 and S2, and then determines that the final user's blood pressure measurement result is 70%*S1+30%. *S2.
- the present application further provides a blood pressure measuring device, which can be used to perform the method embodiment corresponding to FIG. 3 above. Therefore, the embodiment of the blood pressure measuring device provided by the embodiment of the present application can refer to the implementation manner of the method. , the repetition will not be repeated.
- the present application provides a blood pressure measuring device, wherein the collector 210 is configured to collect a biosignal of a user;
- a memory 220 configured to store a computer program
- the processor 230 is configured to call a computer program stored in the memory to perform the following processing:
- Extracting the feature data of the biometric signal according to the biosignal of the user collected by the collector 210; calculating the classification reference value according to the feature data and the preset classification parameter; and predicting the model from the blood pressure according to the preset threshold range in which the classification reference value falls Determining a target blood pressure prediction model in the collection, each blood pressure preset model corresponding to a preset threshold range, and the preset threshold range corresponding to the target blood pressure prediction model is the same as the preset threshold range in which the classification reference value falls; and then inputting the characteristic data into the target
- the blood pressure prediction model obtains the blood pressure measurement result of the user.
- the processor 230 is further configured to: acquire a training data set, where each training data in the training data set includes one feature data and one actual blood pressure measurement result, and the feature data in each training data The corresponding biosignal is the same as the biosignal corresponding to the actual blood pressure measurement result; according to the actual blood pressure measurement result included in each training data, the training data set is divided into at least two first type training data subsets; according to at least two A type of training data subset and a preset classification function obtain preset classification parameters.
- the processor 230 is further configured to: establish a blood pressure prediction model according to each of the first types of training data subsets, and obtain a blood pressure prediction model set.
- the processor 230 is further configured to: divide the training data set into at least two second types of training data subsets according to actual blood pressure measurement results included in each training data, and the first type The number of training data subsets is the same as the number of the second type training data subsets; a blood pressure prediction model is established according to each second type training data subset to obtain a blood pressure prediction model set.
- the number of the first type of training data subset is the same as the number of the second type training data subset, and each of the first type training data subsets corresponds to a first type of actual blood pressure measurement result.
- each second type of training data subset corresponds to a second type of actual blood pressure measurement result range
- the i-th first type of training data subset corresponding to the i-th first type of actual blood pressure measurement result range is A subset of the range of values of the i-th second type of actual blood pressure measurement corresponding to the i-th second-type training data subset, wherein i is a positive integer.
- the steps in the method of the embodiment corresponding to FIG. 3 may be completed by an integrated logic circuit of hardware in the processor 230 or an instruction in a form of software.
- the steps of the blood pressure measurement method disclosed in the embodiments of the present application may be directly implemented as hardware processor execution completion, or performed by a combination of hardware and software modules in the processor.
- Software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable Programmable memories, registers, etc. are well-established in the storage medium of the art.
- the storage medium is located in the memory 220, and the processor 230 reads the information in the memory 220, in conjunction with its hardware, to perform the steps in the method of the embodiment corresponding to FIG. To avoid repetition, it will not be described in detail here.
- the present application also provides a computer readable storage medium having stored therein instructions that, when executed on a computer, cause the computer to perform the method of the embodiment corresponding to FIG. 3 described above.
- the measuring device collects the user's biological signal, extracts the feature data from the biological signal, calculates the classification reference value according to the preset classification parameter and the feature data extracted from the biological signal, and determines the target blood pressure prediction model according to the classification reference value.
- the target blood pressure prediction model determined by the measuring device is also different, and the characteristic data is input to the target blood pressure prediction model to obtain the blood pressure measurement result. Therefore, the blood pressure measurement method provided by the embodiment of the present application can better adapt to the scene of blood pressure diversity, and no longer directly input the feature data into a single blood pressure prediction model, thereby effectively improving the accuracy of the blood pressure measurement, and the method is simple. ,flexible.
- embodiments of the present application can be provided as a method, system, or computer program product. Therefore, the embodiments of the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Moreover, embodiments of the present application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.
- computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
- Embodiments of the present application are described with reference to flowchart illustrations and/or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present application. It will be understood that each flow and/or block of the flowchart illustrations and/or FIG.
- These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine for the execution of instructions for execution by a processor of a computer or other programmable data processing device.
- the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
- the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
- These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
- the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Cardiology (AREA)
- Engineering & Computer Science (AREA)
- Heart & Thoracic Surgery (AREA)
- Physiology (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Vascular Medicine (AREA)
- Biomedical Technology (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
一种血压测量方法及设备,该方法包括:测量设备根据采集到的用户的生物信号,提取生物信号中的特征数据,根据特征数据和预设分类参数计算分类参考值,根据分类参考值落入的预设阈值范围,从血压预测模型集合中确定目标血压预测模型,将特征数据输入目标血压预测模型,获得用户的血压测量结果。因此,本申请提供的血压测量方法能够更好地适应血压多样性的场景,不再直接将特征数据输入到单一的血压预测模型,可以有效提升血压测量的准确度。
Description
本申请要求在2017年1月10日提交中国专利局、申请号为201710017279.8、发明名称为“一种基于分类的血压测量方法和终端”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本申请涉及终端领域,尤其涉及一种血压测量方法及设备。
血压是推动血液在血管内循环流动的动力,能够为各个组织器官提供足够的血量,以维持正常的新陈代谢。其中,血压增高表现出的高血压,是一种很常见的心血管疾病,我国约有1.6亿高血压患者,高血压会带来脑卒、失明、心肌梗死等诸多危害。
近年来,随着移动健康业务的发展,血压监测的便捷性正在逐渐提升,手表、手环等较为方便的可穿戴设备已经能够实现较为便利的测量血压功能。如何提高这些可穿戴设备血压测量的准确度成为一个重要问题。
具体的,可穿戴设备,是基于获取的生物信号来测量血压的,其工作原理为:
1)获取生物信号,通过可穿戴设备中的生物信号采集模块,获取用户的相关生物信号,如心电、脉搏波等生物信号;
2)建立预测模型,根据先验的计算血压的理论公式,或者预先获取的数据,建立血压预测的模型;
3)提取特征数据,对采集到的生物信号进行滤波等预处理操作,然后从执行预处理操作后的生物信号中提取能够表征生物信号的相关特征数据;
4)将提取出的特征数据输入到2)中的预测模型,输出预测的血压值。
由上可知,现有的可穿戴血压设备在测量血压时,都是将生物信号进行预处理后获得相应的特征数据,然后直接将特征数据输入到单一的预测模型中进行血压预测,但是单一的预测模型并不能适应于多样性的血压情况,无法实现高精度预测血压的需求,导致血压测量准确度不高。
发明内容
本申请实施例提供一种血压测量方法及设备,用于解决血压测量准确度不高的问题。
第一方面,本申请提供一种血压测量方法,包括:测量设备根据采集到的用户的生物信号,提取生物信号中的特征数据,根据特征数据和预设分类参数计算分类参考值,然后根据分类参考值落入的预设阈值范围,从血压预测模型集合中确定目标血压预测模型,最后将特征数据输入目标血压预测模型,获得用户的血压测量结果。其中,每个血压预设模型对应一个预设阈值范围,目标血压预测模型对应的预设阈值范围与分类参考值落入的预设阈值范围相同。因此,采用本申请提供的方法测量设备采集用户的生物信号,从生物信号中提取特征数据,根据预设分类参数和从生物信号中提取的特征数据计算分类参考值,
当分类参考值落入的预设阈值范围不同时,测量设备确定的目标血压预测模型也不同,因此,本申请提供的血压测量方法能够更好地适应血压多样性的场景,不再直接将特征数据输入到单一的血压预测模型,进而可以有效提升血压测量的准确度。
在一种可能的实现方式中,预设分类参数是采用以下方法获得的:测量设备获取训练数据集合,根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第一类训练数据子集,根据至少两个第一类训练数据子集和预设分类函数获得预设分类参数,例如,预设的分类函数可以是逻辑回归函数。其中,训练数据集合中的每条训练数据包括一个特征数据和一个实际血压测量结果,实际血压测量结果包括收缩压和舒张压,且每条训练数据中特征数据对应的生物信号与实际血压测量结果对应的生物信号相同。应理解的是,训练数据集合中包括大量训练数据,第k条训练数据中的特征数据是针对第k个生物信号提取出的特征数据,第k条训练数据中的实际血压测量结果是针对第k个生物信号的血压测量结果,这里的实际血压测量结果可以是指用户通过使用水银血压计、有袖带的血压计等工具测量得到的血压测量结果,k为正整数。应理解的是,当预设阈值范围包括两个以上,则需要两组以上预设分类参数,需要计算得到的两个以上分类参考值进行组合,以分别对应不同的预设阈值范围。
在一种可能的实现方式中,血压预测模型集合是采用以下方法获得的:
测量设备根据每个第一类训练数据子集建立一个血压预测模型,获得血压预测模型集合。因此,测量设备可以直接根据第一类训练数据子集建立血压预测模型,方法简便。
在一种可能的实现方式中,血压预测模型集合是采用以下方法获得的:
测量设备根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第二类训练数据子集,第一类训练数据子集的数目与第二类训练数据子集的数目相同;根据每个第二类训练数据子集建立一个血压预测模型,获得血压预测模型集合。因此,测量设备可以采用与第一类训练数据子集不同的第二类训练数据子集建立血压预测模型,方法灵活,以适用于不同场景的需要。
在一种可能的实现方式中,第一类训练数据子集的数目与第二类训练数据子集的数目相同,每个第一类训练数据子集对应一个第一类实际血压测量结果取值范围,每个第二类训练数据子集对应一个第二类实际血压测量结果取值范围,第i个第一类训练数据子集对应的第i个第一类实际血压测量结果取值范围是第i个第二类训练数据子集对应的第i个第二类实际血压测量结果取值范围的子集,其中,i为正整数。因此,将训练数据集合划分为至少两个第二类训练数据子集,而不再继续使用至少两个第一类训练数据子集的原因在于,建立不同血压预测模型所采用的训练数据子集之间需要有一定的冗余度,以保证分类节点附近的血压测量的准确性。
第二方面,本申请提供一种血压测量设备,包括:采集器,存储器和处理器,其中,采集器,用于采集用户的生物信号;存储器,用于存储计算机程序;处理器,用于调用存储器中存储的计算机程序,执行:根据采集器采集到的用户的生物信号,提取生物信号中的特征数据;根据特征数据和预设分类参数计算分类参考值;根据分类参考值落入的预设阈值范围,从血压预测模型集合中确定目标血压预测模型,每个血压预设模型对应一个预设阈值范围,目标血压预测模型对应的预设阈值范围与分类参考值落入的预设阈值范围相同;将特征数据输入目标血压预测模型,获得用户的血压测量结果。
在一种可能的实现方式中,处理器,还用于:获取训练数据集合,训练数据集合中的
每条训练数据包括一个特征数据和一个实际血压测量结果,且每条训练数据中特征数据对应的生物信号与实际血压测量结果对应的生物信号相同;根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第一类训练数据子集;根据至少两个第一类训练数据子集和预设分类函数获得预设分类参数。
在一种可能的实现方式中,处理器,还用于:根据每个第一类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
在一种可能的实现方式中,处理器,还用于:根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第二类训练数据子集,第一类训练数据子集的数目与第二类训练数据子集的数目相同;根据每个第二类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
在一种可能的实现方式中,第一类训练数据子集的数目与第二类训练数据子集的数目相同,每个第一类训练数据子集对应一个第一类实际血压测量结果取值范围,每个第二类训练数据子集对应一个第二类实际血压测量结果取值范围,第i个第一类训练数据子集对应的第i个第一类实际血压测量结果取值范围是第i个第二类训练数据子集对应的第i个第二类实际血压测量结果取值范围的子集,其中,i为正整数。
第三方面,本申请提供了一种计算机可读存储介质,所述计算机可读存储介质中存储有指令,当所述指令在计算机上运行时,使得计算机执行上述各方面所述的方法。
第四方面,本申请提供了一种包含指令的计算机程序产品,当所述计算机程序产品在计算机上运行时,使得计算机执行上述各方面所述的方法。
图1为本申请实施例中具有血压测量功能的手表的示意图;
图2为本申请实施例中测量设备的基本构成示意图;
图3为本申请实施例中血压测量方法的概述流程图;
图4为本申请实施例中测量设备执行血压测量的具体流程图;
图5为本申请实施例中训练数据集合划分为两个第二类训练数据子集的数据分布示意图;
图6为本申请实施例中根据第二类训练数据子集划分结果获得对应血压预测模型的示意图。
下面结合附图,对本申请的实施例进行描述。
应理解的是,本申请实施例中提到的测量设备一般是指可穿戴设备,具体的,可穿戴设备可以为手表、手环等,如图1所示为一种具有血压测量功能的手表。
图2所示为测量设备的基本构成示意图,该测量设备包括:采集器210,存储器220,处理器230,电源240等。
在一种可能的设计中,测量设备还可包括:显示单元250,蓝牙模块260等。
本申请实施例中提到的测量设备与上述图2所示的测量设备构成基本一致,须知本申请实施例不涉及对测量设备硬件构成的改进。
参阅图3所示,本申请实施例提供一种血压测量方法,以解决血压测量准确度不高的问题,该方法包括:
步骤300:测量设备根据采集到的用户的生物信号,提取生物信号中的特征数据。
这里的生物信号可以为心电信号,或脉搏波信号等。
从生物信号中提取的特征数据可以为波峰波谷之间的间隔、波峰的半高全宽等能够表征生物信号的特征数据。提取生物信号中特征数据的一般的做法是先提取生物信号中的特征点的位置,如脉搏波信号的波峰、波谷和最大斜率点等位置,然后基于特征点的位置,结合生物信号获得特征数据。例如,从生物信号中提取的特征数据可以记为x0,x1,x2,…,xn。
步骤310:测量设备根据特征数据和预设分类参数计算分类参考值。
这里的预设分类参数可以为A,记为a0,a1,a2,…,an。例如,当特征数据为x0,x1,x2,…,xn时,测量设备计算得到的分类参考值为a0*x0+a1*x1+a2*x2+…+an*xn。
步骤320:测量设备根据分类参考值落入的预设阈值范围,从血压预测模型集合中确定目标血压预测模型。
每个血压预设模型对应一个预设阈值范围,目标血压预测模型对应的预设阈值范围与分类参考值落入的预设阈值范围相同。
例如,假设预设阈值范围包括两个,分别为分类参考值大于等于0和分类参考值小于0,则血压预测模型集合也包括两个,分别对应分类参考值大于等于0和分类参考值小于0。当步骤310计算得到的分类参考值大于等于0时,测量设备将分类参考值大于等于0对应的血压预测模型作为目标血压预测模型,当步骤310计算得到的分类参考值小于0时,测量设备将分类参考值小于0对应的血压预测模型作为目标血压预测模型。
应理解的是,当预设阈值范围包括两个以上,则需要两组以上预设分类参数,需要计算得到的两个以上分类参考值进行组合,以分别对应不同的预设阈值范围。
步骤330:测量设备将特征数据输入目标血压预测模型,获得用户的血压测量结果。
因此,通过上述步骤300~步骤320,测量设备确定出目标血压预测模型,即适用于当前特征数据的血压预设模型,进而将特征数据输入至该目标血压预测模型。
须知,血压预测模型集合中的每个血压预测模型是以特征数据为自变量,以血压值为因变量。
由上可知,如图4所示,测量设备采集用户的生物信号,从生物信号中提取特征数据,根据预设分类参数和从生物信号中提取的特征数据计算分类参考值,根据分类参考值确定目标血压预测模型,将特征数据输入至目标血压预测模型获得血压测量结果。当分类参考值落入的预设阈值范围不同时,测量设备确定的目标血压预测模型也不同,因此,本申请实施例提供的血压测量方法能够更好地适应血压多样性的场景,不再直接将特征数据输入到单一的血压预测模型,进而有效提升了血压测量的准确度。
应理解的是,步骤310中提到的预设分类参数和步骤320中提到的血压预测模型集合是在测量设备执行血压测量之前获得,保存在测量设备中。
上述预设分类参数和血压预测模型集合可以通过建立分类模型和血压预测模型获得,具体可以采用但不限于以下可能的实现方式。
首先,测量设备获取训练数据集合。
其中,训练数据集合中的每条训练数据包括一个特征数据和一个实际血压测量结果,实际血压测量结果包括收缩压和舒张压,且每条训练数据中特征数据对应的生物信号与实
际血压测量结果对应的生物信号相同。
应理解的是,训练数据集合中包括大量训练数据,第k条训练数据中的特征数据是针对第k个生物信号提取出的特征数据,第k条训练数据中的实际血压测量结果是针对第k个生物信号的血压测量结果,这里的实际血压测量结果可以是指用户通过使用水银血压计、有袖带的血压计等工具测量得到的血压测量结果,k为正整数。
然后,测量设备根据训练数据集合建立分类模型,获得预设分类参数。具体的,测量设备根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第一类训练数据子集,根据至少两个第一类训练数据子集和预设分类函数,获得预设分类参数。其中,每个第一类训练数据子集对应一个第一类实际血压测量结果取值范围。
例如,以收缩压(systolic blood pressure,SBP)等于140为界,将训练数据集合分为两个第一类训练数据子集,第一个第一类训练数据子集包括训练数据中SBP小于140的训练数据,第二个第一类训练数据子集包括训练数据中SBP大于等于140的训练数据。其中,将训练数据中SBP小于140的训练数据标记为-1,训练数据中SBP大于等于140的训练数据标记为+1,通过预设的分类函数获得预设分类参数A,记为a0,a1,a2,…,an。
应理解的是,上述以SBP等于140为界,将训练数据集合划分到两个第一类训练数据子集,其中,第一个第一类训练数据子集表示非高压子集,第二个第一类训练数据子集表示高压子集,仅作为一种举例,也可以根据实际情况进行其他分类。例如,以SBP等于100为界,将训练数据集合划分到两个第一类训练数据子集,第一个第一类训练数据子集包括训练数据中SBP小于100的训练数据,表示低压子集,第二个第一类训练数据子集包括训练数据中SBP大于等于100的训练数据,表示非低压子集。当然,也可以以SBP等于110、120等值为界,将训练数据集合划分为任意两个第一类训练数据子集,不考虑其中是否具有高压、低压等特定含义。
接着,测量设备根据每个第一类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
例如,测量设备以SBP等于140为界,将训练数据集合分为两个第一类训练数据子集,已根据这两个第一类训练数据子集和预设分类函数获得预设分类参数。进一步地,测量设备根据这两个第一类训练数据子集通过预设回归函数获得两个血压预测模型,分别记为模型1和模型2,例如,回归函数可以是线性回归函数,其中,模型1的参数B,记为b0,b1,b2,…,bn,模型1对应的分类参考值小于0,模型2的参数C,记为c0,c1,c2,…,cn,模型2对应的分类参考值大于等于0。
进一步地,假设特征数据为x0,x1,x2,…,xn,当测量设备计算得到的分类参考值小于0时,测量设备将模型1作为目标血压预测模型,获得用户的血压测量结果为b0*x0+b1*x1+b2*x2+…+bn*xn;当测量设备计算得到的分类参考值大于等于0时,测量设备将模型2作为目标血压预测模型,获得用户的血压测量结果为c0*x0+c1*x1+c2*x2+…+cn*xn。
应理解的是,实施例中是以收缩压为例进行阐述的,预测舒张压的方式与预测收缩压的方式类似,不同之处在于将建立模型时的因变量由收缩压改为舒张压。
可选的,获得血压预测模型集合还可采用如下方案:
测量设备根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第二类训练数据子集,根据每个第二类训练数据子集建立一个血压预测模型,获得血
压预测模型集合。其中,每个第二类训练数据子集对应一个第二类实际血压测量结果取值范围。
可选的,第一类训练数据子集的数目与第二类训练数据子集的数目相同,第i个第一类训练数据子集对应的第i个第一类实际血压测量结果取值范围是第i个第二类训练数据子集对应的第i个第二类实际血压测量结果取值范围的子集,其中,i为正整数。
例如,测量设备根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为两个第二类训练数据子集,如图5所示,第一个第二类训练数据子集包括训练数据中SBP小于150的训练数据,第二个第二类训练数据子集包括训练数据中SBP大于等于130的训练数据,然后,测量设备根据这两个第二类训练数据子集通过预设回归函数获得两个血压预测模型,分别为模型x和模型y,如图6所述,其中,模型x对应第一个第一类训练数据子集,模型x的参数B,记为b0,b1,b2,…,bn,模型x还对应分类参考值小于0,模型y对应第二个第一类训练数据子集,模型y的参数C,记为c0,c1,c2,…,cn,模型y还对应分类参考值大于等于0。
在上述实施例中,将训练数据集合划分为至少两个第二类训练数据子集,而不再继续使用至少两个第一类训练数据子集的原因在于,建立不同血压预测模型所采用的训练数据子集之间需要有一定的冗余度,以保证分类节点附近的血压测量的准确性。
又例如,以SBP等于100和SBP等于140为界将训练数据集合划分成四个第一类训练数据子集,第一个第一类训练数据子集包括训练数据中SBP小于100的训练数据,第二个第一类训练数据子集包括训练数据中SBP大于等于100的训练数据,第三个第一类训练数据子集包括训练数据中SBP小于140的训练数据,第四个第一类训练数据子集包括训练数据中SBP大于等于140的训练数据,基于第一个第一类训练数据子集和第二个第一类训练数据子集,通过预设的分类函数获得预设分类参数A1,基于第三个第一类训练数据子集和第四个第一类训练数据子集,通过预设的分类函数获得预设分类参数A2。
进一步地,将训练数据集合划分为三个第二类训练数据子集,第一个第二类训练数据子集包括训练数据中SBP小于100的训练数据,第二个第一类训练数据子集包括训练数据中SBP大于等于100且SBP小于140的训练数据,第三个第一类训练数据子集包括训练数据中SBP大于等于140的训练数据,基于这三个第二类训练数据子集通过预设回归函数获得三个血压预测模型,分别为模型x1、模型x2和模型x3。此外,在将训练数据集合划分为三个第二类训练数据子集时,也可以为各个训练数据子集之间设定一定的冗余度。
其中,模型x1对应根据A1计算得到分类参考值小于0,模型x2对应根据A1计算的分类参考值大于等于0且根据A2计算得到分类参考值小于0,模型x3对应根据A2计算得到分类参考值大于等于0。
当特征数据为x0,x1,x2,…,xn时,测量设备根据A1计算分类参考值,记为B1,根据A2计算分类参考值,记为B2,当B1小于0时,测量设备将x1作为目标血压预测模型,当B1大于等于0且B2小于0时,测量设备将x2作为目标血压预测模型,当B3大于等于0时,测量设备将x3作为目标血压预测模型。
应理解的是,上述获得预设分类参数和血压预测模型集合的过程可以由测量设备执行,也可由其他设备执行,在其他设备执行上述过程获得预设分类参数和血压预测模型集合后,需将预设分类参数和血压预测模型集合存储至测量设备中。
此外,可选的,测量设备根据特征数据和预设分类参数计算分类参考值后,测量设备
根据分类参考值落入的预设阈值范围,还可以从血压预测模型集合中确定至少两个目标血压预测模型,所述至少两个目标血压预测模型的概率和为1,测量设备将特征数据输入至每个目标血压预测模型,获得所述至少两个目标血压预测模型分别对应血压测量结果,然后根据每个目标血压预测模型的概率确定最终的用户的血压测量结果。
例如,测量设备根据分类参考值落入的预设阈值范围,从血压预测模型集合中确定两个目标血压预测模型,分别为模型1和模型2,模型1的概率为70%,模型2的概率为30%,测量设备将特征数据分别输入至这两个目标血压预测模型,获得两个血压测量结果,分别为S1和S2,然后确定最终的用户的血压测量结果为70%*S1+30%*S2。
基于同一构思,本申请还提供了一种血压测量设备,该设备可以用于执行上述图3对应的方法实施例,因此本申请实施例提供的血压测量设备的实施方式可以参见该方法的实施方式,重复之处不再赘述。
如图2所示,本申请提供一种血压测量设备,其中,采集器210,用于采集用户的生物信号;
存储器220,用于存储计算机程序;
处理器230,用于调用存储器中存储的计算机程序,以执行如下处理:
根据采集器210采集到的用户的生物信号,提取生物特征信号的特征数据;根据特征数据和预设分类参数计算分类参考值;并根据分类参考值落入的预设阈值范围,从血压预测模型集合中确定目标血压预测模型,每个血压预设模型对应一个预设阈值范围,目标血压预测模型对应的预设阈值范围与分类参考值落入的预设阈值范围相同;然后将特征数据输入目标血压预测模型,获得用户的血压测量结果。
在一种可能的实现方式中,处理器230,还用于:获取训练数据集合,训练数据集合中的每条训练数据包括一个特征数据和一个实际血压测量结果,且每条训练数据中特征数据对应的生物信号与实际血压测量结果对应的生物信号相同;根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第一类训练数据子集;根据至少两个第一类训练数据子集和预设分类函数获得预设分类参数。
在一种可能的实现方式中,处理器230,还用于:根据每个第一类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
在一种可能的实现方式中,处理器230,还用于:根据每条训练数据中包括的实际血压测量结果,将训练数据集合划分为至少两个第二类训练数据子集,第一类训练数据子集的数目与第二类训练数据子集的数目相同;根据每个第二类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
在一种可能的实现方式中,第一类训练数据子集的数目与第二类训练数据子集的数目相同,每个第一类训练数据子集对应一个第一类实际血压测量结果取值范围,每个第二类训练数据子集对应一个第二类实际血压测量结果取值范围,第i个第一类训练数据子集对应的第i个第一类实际血压测量结果取值范围是第i个第二类训练数据子集对应的第i个第二类实际血压测量结果取值范围的子集,其中,i为正整数。
在具体实现过程中,图3对应的实施例的方法中的步骤可以通过处理器230中的硬件的集成逻辑电路或者软件形式的指令完成。结合本申请实施例所公开的血压测量方法的步骤可以直接体现为硬件处理器执行完成,或者用处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器,闪存、只读存储器,可编程只读存储器或者电可擦写
可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器220中,处理器230读取存储器220中的信息,结合其硬件完成图3对应的实施例的方法中的步骤。为避免重复,这里不再详细描述。
本申请还提供了一种计算机可读存储介质,该计算机可读存储介质中存储有指令,当指令在计算机上运行时,使得计算机可以执行上述如图3对应的实施例的方法。
综上所述,测量设备采集用户的生物信号,从生物信号中提取特征数据,根据预设分类参数和从生物信号中提取的特征数据计算分类参考值,根据分类参考值确定目标血压预测模型,当分类参考值落入的预设阈值范围不同时,测量设备确定的目标血压预测模型也不同,将特征数据输入至目标血压预测模型获得血压测量结果。因此,本申请实施例提供的血压测量方法能够更好地适应血压多样性的场景,不再直接将特征数据输入到单一的血压预测模型,进而可以有效提升了血压测量的准确度,且方法简便、灵活。此外,建立不同血压预测模型所采用的训练数据子集之间具有一定的冗余度,以保证分类节点附近的血压测量的准确性。
本领域内的技术人员应明白,本申请实施例可提供为方法、***、或计算机程序产品。因此,本申请实施例可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本申请实施例可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请实施例是参照根据本申请实施例的方法、设备(***)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然,本领域的技术人员可以对本申请实施例进行各种改动和变型而不脱离本申请的精神和范围。这样,倘若本申请实施例的这些修改和变型属于本申请权利要求及其等同技术的范围之内,则本申请也意图包含这些改动和变型在内。
Claims (11)
- 一种血压测量方法,其特征在于,包括:测量设备根据采集到的用户的生物信号,提取所述生物信号中的特征数据;所述测量设备根据所述特征数据和预设分类参数计算分类参考值;所述测量设备根据所述分类参考值落入的预设阈值范围,从血压预测模型集合中确定目标血压预测模型,每个血压预设模型对应一个预设阈值范围,所述目标血压预测模型对应的预设阈值范围与所述分类参考值落入的预设阈值范围相同;所述测量设备将所述特征数据输入所述目标血压预测模型,获得所述用户的血压测量结果。
- 如权利要求1所述的方法,其特征在于,所述预设分类参数是采用以下方法获得的:所述测量设备获取训练数据集合,所述训练数据集合中的每条训练数据包括一个特征数据和一个实际血压测量结果,且每条训练数据中特征数据对应的生物信号与实际血压测量结果对应的生物信号相同;所述测量设备根据每条训练数据中包括的实际血压测量结果,将所述训练数据集合划分为至少两个第一类训练数据子集;所述测量设备根据所述至少两个第一类训练数据子集和预设分类函数获得预设分类参数。
- 如权利要求2所述的方法,其特征在于,所述血压预测模型集合是采用以下方法获得的:所述测量设备根据每个第一类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
- 如权利要求2所述的方法,其特征在于,所述血压预测模型集合是采用以下方法获得的:所述测量设备根据每条训练数据中包括的实际血压测量结果,将所述训练数据集合划分为至少两个第二类训练数据子集,第一类训练数据子集的数目与第二类训练数据子集的数目相同;所述测量设备根据每个第二类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
- 如权利要求4所述的方法,其特征在于,第一类训练数据子集的数目与第二类训练数据子集的数目相同,每个第一类训练数据子集对应一个第一类实际血压测量结果取值范围,每个第二类训练数据子集对应一个第二类实际血压测量结果取值范围,第i个第一类训练数据子集对应的第i个第一类实际血压测量结果取值范围是第i个第二类训练数据子集对应的第i个第二类实际血压测量结果取值范围的子集,其中,i为正整数。
- 一种血压测量设备,其特征在于,包括:采集器、存储器和处理器,其中:所述采集器,用于采集用户的生物信号;所述存储器,用于存储计算机程序;所述处理器,用于调用所述存储器中存储的计算机程序,执行:根据所述采集器采集到的用户的生物信号,提取所述生物信号中的特征数据;根据所述特征数据和预设分类参 数计算分类参考值;根据所述分类参考值落入的预设阈值范围,从血压预测模型集合中确定目标血压预测模型,每个血压预设模型对应一个预设阈值范围,所述目标血压预测模型对应的预设阈值范围与所述分类参考值落入的预设阈值范围相同;将所述特征数据输入所述目标血压预测模型,获得所述用户的血压测量结果。
- 如权利要求6所述的设备,其特征在于,所述处理器,还用于:获取训练数据集合,所述训练数据集合中的每条训练数据包括一个特征数据和一个实际血压测量结果,且每条训练数据中特征数据对应的生物信号与实际血压测量结果对应的生物信号相同;根据每条训练数据中包括的实际血压测量结果,将所述训练数据集合划分为至少两个第一类训练数据子集;根据所述至少两个第一类训练数据子集和预设分类函数获得预设分类参数。
- 如权利要求7所述的设备,其特征在于,所述处理器,还用于:根据每个第一类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
- 如权利要求8所述的设备,其特征在于,所述处理器,还用于:根据每条训练数据中包括的实际血压测量结果,将所述训练数据集合划分为至少两个第二类训练数据子集,第一类训练数据子集的数目与第二类训练数据子集的数目相同;根据每个第二类训练数据子集建立一个血压预测模型,获得血压预测模型集合。
- 如权利要求9所述的设备,其特征在于,第一类训练数据子集的数目与第二类训练数据子集的数目相同,每个第一类训练数据子集对应一个第一类实际血压测量结果取值范围,每个第二类训练数据子集对应一个第二类实际血压测量结果取值范围,第i个第一类训练数据子集对应的第i个第一类实际血压测量结果取值范围是第i个第二类训练数据子集对应的第i个第二类实际血压测量结果取值范围的子集,其中,i为正整数。
- 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质中存储有指令,当所述指令在计算机上运行时,使得计算机执行如权利要求1~5任一项所述的方法。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201780009087.0A CN108697350A (zh) | 2017-01-10 | 2017-04-17 | 一种血压测量方法及设备 |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710017279 | 2017-01-10 | ||
CN201710017279.8 | 2017-01-10 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2018129822A1 true WO2018129822A1 (zh) | 2018-07-19 |
Family
ID=62839763
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2017/080811 WO2018129822A1 (zh) | 2017-01-10 | 2017-04-17 | 一种血压测量方法及设备 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN108697350A (zh) |
WO (1) | WO2018129822A1 (zh) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110956480A (zh) * | 2018-09-26 | 2020-04-03 | 北京嘀嘀无限科技发展有限公司 | 用户结构预估方法、装置及服务器 |
CN113057611A (zh) * | 2021-03-19 | 2021-07-02 | 北京京东拓先科技有限公司 | 用于输出信息的方法、装置、设备以及存储介质 |
CN113288091A (zh) * | 2021-05-06 | 2021-08-24 | 广东工业大学 | 用于血压分类的模型训练方法、装置及可穿戴设备 |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109948280B (zh) * | 2019-03-29 | 2023-06-09 | 广州视源电子科技股份有限公司 | 无袖带血压模型生成方法、装置、设备及可读存储介质 |
CN113491513B (zh) * | 2020-04-08 | 2023-06-30 | 华为技术有限公司 | 一种心律检测控制方法及终端 |
CN114052685B (zh) * | 2020-08-06 | 2023-10-13 | 华为技术有限公司 | 电子设备及计算机可读存储介质 |
CN114021076A (zh) * | 2021-03-19 | 2022-02-08 | 北京超思电子技术有限责任公司 | 基于ecg和ppg信号的血压计算模型生成方法及血压测量*** |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110066051A1 (en) * | 2009-09-15 | 2011-03-17 | Jim Moon | Body-worn vital sign monitor |
CN102894964A (zh) * | 2011-07-26 | 2013-01-30 | 深圳大学 | 一种无创血压测量方法和装置 |
CN103976721A (zh) * | 2014-04-22 | 2014-08-13 | 辛勤 | 血压测量方法以及用于实现该方法的嵌入式装置 |
CN104739395A (zh) * | 2015-03-25 | 2015-07-01 | 华中科技大学 | 一种基于脉搏波的人体血压预测方法 |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105748051B (zh) * | 2016-02-18 | 2018-10-09 | 京东方科技集团股份有限公司 | 一种血压测量装置 |
-
2017
- 2017-04-17 WO PCT/CN2017/080811 patent/WO2018129822A1/zh active Application Filing
- 2017-04-17 CN CN201780009087.0A patent/CN108697350A/zh active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20110066051A1 (en) * | 2009-09-15 | 2011-03-17 | Jim Moon | Body-worn vital sign monitor |
CN102894964A (zh) * | 2011-07-26 | 2013-01-30 | 深圳大学 | 一种无创血压测量方法和装置 |
CN103976721A (zh) * | 2014-04-22 | 2014-08-13 | 辛勤 | 血压测量方法以及用于实现该方法的嵌入式装置 |
CN104739395A (zh) * | 2015-03-25 | 2015-07-01 | 华中科技大学 | 一种基于脉搏波的人体血压预测方法 |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110956480A (zh) * | 2018-09-26 | 2020-04-03 | 北京嘀嘀无限科技发展有限公司 | 用户结构预估方法、装置及服务器 |
CN113057611A (zh) * | 2021-03-19 | 2021-07-02 | 北京京东拓先科技有限公司 | 用于输出信息的方法、装置、设备以及存储介质 |
CN113057611B (zh) * | 2021-03-19 | 2024-01-12 | 北京京东拓先科技有限公司 | 用于输出信息的方法、装置、设备以及存储介质 |
CN113288091A (zh) * | 2021-05-06 | 2021-08-24 | 广东工业大学 | 用于血压分类的模型训练方法、装置及可穿戴设备 |
CN113288091B (zh) * | 2021-05-06 | 2023-10-03 | 广东工业大学 | 用于血压分类的模型训练方法、装置及可穿戴设备 |
Also Published As
Publication number | Publication date |
---|---|
CN108697350A (zh) | 2018-10-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2018129822A1 (zh) | 一种血压测量方法及设备 | |
US9161705B2 (en) | Method and device for early detection of heart attack | |
US10980429B2 (en) | Method and system for cuffless blood pressure estimation using photoplethysmogram features and pulse transit time | |
WO2018166136A1 (zh) | 波形信号处理的方法及装置 | |
CN110755059B (zh) | 一种血压波形监测方法及装置 | |
CN102551701A (zh) | 通过周期性信号分析检测实体物件异常运作的***及方法 | |
CN109009084B (zh) | 多导联心电信号的qrs波群校验方法、装置及设备、介质 | |
US20220254028A1 (en) | Method and apparatus for adjusting blood flow velocity in maximum hyperemia state based on index for microcirculatory resistance | |
JP6583892B2 (ja) | Ecg信号の並行解析装置、方法及び移動端末 | |
CN109512395B (zh) | 一种生物信号的分析处理方法、装置及设备 | |
US20220261997A1 (en) | Method and apparatus for acquiring blood vessel evaluation parameter based on physiological parameter, and storage medium | |
Antink et al. | Detection of heart beats in multimodal data: a robust beat-to-beat interval estimation approach | |
US20240164692A1 (en) | Electrocardiogram analysis assistance device, program, electrocardiogram analysis assistance method, electrocardiogram analysis assistance system, peak estimation model generation method, and segment estimation model generation method | |
CN110432894B (zh) | 心电图关键点标注方法及电子设备 | |
US20240164691A1 (en) | Electrocardiogram analysis assistance device, program, electrocardiogram analysis assistance method, and electrocardiogram analysis assistance system | |
JP2019069044A (ja) | 瞬時心拍の時系列データの補完装置、補完方法及びそのプログラム | |
CN113116358B (zh) | 心电图的显示方法、装置、终端设备和存储介质 | |
WO2021068193A1 (zh) | 一种血压波形监测方法及装置 | |
CN113768476B (zh) | 心动周期时相检测方法、装置及计算机程序产品 | |
Risman et al. | Implementation of heart rate variability analysis algorithm on FPGA platform | |
CN116807490A (zh) | 心电图的处理方法、装置、医疗设备及存储介质 | |
WO2018023696A1 (zh) | 一种胎儿心电分离方法及装置 | |
CN114668401B (zh) | 一种ai心电训练数据标注方法、装置、电子设备及介质 | |
KR102257498B1 (ko) | 활동 전위 기간 상환 현상을 반영한 심장 모델 생성 방법 및 생성 장치 | |
US20230148879A1 (en) | Computer-based platforms and systems configured for cuff-less blood pressure estimation from photoplethysmography via visibility graph and transfer learning and methods of use thereof |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 17891244 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 17891244 Country of ref document: EP Kind code of ref document: A1 |