CN109791564B - Method and device for setting parameters in signal calculation method - Google Patents
Method and device for setting parameters in signal calculation method Download PDFInfo
- Publication number
- CN109791564B CN109791564B CN201780000780.1A CN201780000780A CN109791564B CN 109791564 B CN109791564 B CN 109791564B CN 201780000780 A CN201780000780 A CN 201780000780A CN 109791564 B CN109791564 B CN 109791564B
- Authority
- CN
- China
- Prior art keywords
- parameters
- parameter
- signal
- calculation method
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
- Indication And Recording Devices For Special Purposes And Tariff Metering Devices (AREA)
Abstract
Some embodiments of the application provide a method and a device for setting parameters in a signal calculation method. The setting method of the parameters in the signal calculation method comprises the following steps: establishing a parameter space, wherein the parameter space comprises at least one group of parameters with values within the parameter value range in the signal calculation method; establishing a measurement model for evaluating the quality of the parameters; substituting the parameters in the signal and parameter space in the preset signal data set into a measurement model, and calculating to obtain a parameter quality measurement value; and setting parameters in the signal calculation method according to the parameter quality measurement values. By adopting the embodiment of the application, the quality degree of the parameters in the parameter space is evaluated through the measurement model, the more accurate parameters are objectively set for the signal calculation method, and subjectivity and uncertainty of manually setting the parameters are avoided.
Description
Technical Field
The present invention relates to the field of signal processing technologies, and in particular, to a method and an apparatus for setting parameters in a signal calculation method.
Background
Along with the improvement of living standard, people pay more attention to the health level of life. Heart rate is the number of beats per minute of the human heart, and is a very important physiological index in clinical diagnosis. Traditional medical equipment requires a user to be in a static state when measuring heart rate, and is inconvenient to carry; accordingly, many manufacturers have produced wearable devices that can perform heart rate measurements so that a user can perform heart rate measurements in a state of daily life.
The most commonly used heart rate measurement method in the prior art is a photoplethysmography (PPG) method, wherein an LED is used to emit light with a specific wavelength, and the light is transmitted, scattered, diffracted and reflected by human tissues and then returned, and the returned light signal is converted into an electrical signal, so that a corresponding PPG signal is obtained. The light beam is attenuated in the propagation process of human tissues due to the absorption effect of the human tissues, wherein the absorption of static tissues such as skin, fat, muscle and the like is a constant value, and the blood generates periodical volume change due to the contraction and relaxation cycles of the heart, so that a periodical waveform consistent with the heartbeat is generated in the PPG signal, the heartbeat frequency can be measured through the PPG signal, and the measurement of the heart rate by the photoelectric pulse volume method is a noninvasive and harmless measurement method.
The inventors found that the prior art has at least the following problems: the existing heart rate calculation method mainly comprises a time domain waveform method, a frequency spectrum analysis method, an independent component analysis method, an empirical mode decomposition method, an adaptive filter method and the like, and a plurality of parameter variables are arranged in the heart rate calculation method, wherein the range of the parameter variables can be estimated according to a mathematical principle, the specific numerical value of the parameter variables can be set according to experience or clinical physiological characteristics, but the setting of the parameter variables has subjectivity and uncertainty, and the accuracy of the set parameter variables cannot be ensured.
Disclosure of Invention
The embodiment of the application aims to provide a method and a device for setting parameters in a signal calculation method, wherein the method and the device are used for evaluating the quality degree of the parameters in a parameter space through a measurement model, and objectively setting more accurate parameters for the signal calculation method, so that subjectivity and uncertainty of manually setting the parameters are avoided.
The embodiment of the application provides a method for setting parameters in a signal calculation method, which comprises the following steps: establishing a parameter space, wherein the parameter space comprises at least one group of parameters with values within the parameter value range in the signal calculation method; establishing a measurement model for evaluating the quality of the parameters; substituting the parameters in the signal and parameter space in the preset signal data set into a measurement model, and calculating to obtain a parameter quality measurement value; and setting parameters in the signal calculation method according to the parameter quality measurement values.
The embodiment of the application also provides a device for setting parameters in a signal calculation method, which comprises the following steps: the first establishing module is used for establishing a parameter space; the parameter space comprises at least one group of parameters with values within the parameter value range in the signal calculation method; the second building module is used for building a measurement model for evaluating the quality of the parameters; the calculation module is used for substituting the parameters in the signal and parameter space in the preset signal data set into the measurement model, and calculating to obtain the parameter quality measurement value; and the setting module is used for setting parameters in the signal calculation method according to the parameter quality measurement value.
Compared with the prior art, the embodiment of the application establishes a parameter space and establishes the measurement model for evaluating the quality of the parameters, so that the quality degree of the parameters in the parameter space can be evaluated through the measurement model, more accurate parameters are objectively set for a signal calculation method, and subjectivity and uncertainty of manually setting the parameters are avoided.
In addition, the parameter space includes multiple sets of parameters; substituting the parameters in the signal and parameter space in the preset signal data set into a measurement model, and calculating to obtain a parameter quality measurement value, wherein the parameter quality measurement value specifically comprises: substituting the parameters in the signal and parameter space in the preset signal data set into a measurement model, and calculating to obtain the parameter quality measurement values of a plurality of groups of parameters; parameters in the signal calculation method are set according to the parameter quality measurement values, and specifically: and setting a group of parameters with the smallest parameter quality metric value as parameters in a signal calculation method. The embodiment provides a specific implementation mode for setting parameters in a signal calculation method.
In addition, the parameters in the signal and parameter space in the preset signal data set are substituted into the measurement model, and the parameter quality measurement value is obtained through calculation, specifically: substituting the parameters in the signal and parameter space in the preset signal data set into a measurement model, and calculating by adopting a gradient descent method to obtain a parameter quality measurement value meeting the preset condition; wherein, any group of parameters in the parameter space is used as the initial value of the gradient descent method; parameters in the signal calculation method are set according to the parameter quality measurement values, and specifically: and setting a group of parameters corresponding to the parameter quality metric values meeting the preset conditions as parameters in the signal calculation method. The embodiment provides another specific implementation mode for setting parameters in the signal calculation method.
In addition, the parameters in the signal and parameter space in the preset signal data set are substituted into the measurement model, and the parameter quality measurement values of a plurality of groups of parameters are obtained through calculation, specifically: substituting the parameters in the signal and parameter space in the preset signal data set into the measurement model, and performing parallel calculation to obtain the parameter quality measurement values of multiple groups of parameters. In this embodiment, the parameter quality metric values of multiple groups of parameters are calculated in parallel, so that the calculation speed is increased, and the calculation time is reduced.
In addition, before setting a group of parameters with the smallest parameter quality metric value as parameters in the signal calculation method, the method further comprises: and filtering the parameter quality measurement values of the plurality of groups of parameters. In this embodiment, the filtering process is performed on the parameter quality metrics of multiple sets of parameters, so that errors of the parameter quality metrics caused by noise or abnormal data can be prevented.
In addition, the filtering processing mode is mean smoothing filtering or median filtering. The embodiment provides a specific filtering processing mode.
In addition, the metric model is a cost function based on minimum mean square error or a cost function based on least squares. The present embodiment provides a specific implementation of the metrology model.
Drawings
One or more embodiments are illustrated by way of example and not limitation in the figures of the accompanying drawings, in which like references indicate similar elements, and in which the figures of the drawings are not to be taken in a limiting sense, unless otherwise indicated.
Fig. 1 is a specific flowchart of a setting method of parameters in a signal calculation method according to a first embodiment of the present application;
fig. 2 is a specific flowchart of a method for setting parameters in a signal calculation method according to a second embodiment of the present application;
fig. 3 is a specific flowchart of a method for setting parameters in a signal calculation method according to a third embodiment of the present application;
fig. 4 is a specific flowchart of a setting method of parameters in a signal calculation method according to a fourth embodiment of the present application;
fig. 5 is a block schematic diagram of a setting device of parameters in a signal calculation method according to a fifth embodiment of the present application;
fig. 6 is a block diagram of a setting device of parameters in a signal calculation method according to an eighth embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, some embodiments of the present application will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
The first embodiment of the present application relates to a method for setting parameters in a signal calculation method, where the signal may be a heart rate signal, an oxygen signal, and other biological signals, and the signal calculation method may be an adaptive filtering algorithm for calculating a heart rate, a frequency domain heart rate algorithm for calculating a heart rate, an oxygen calculation algorithm for calculating oxygen, and the like, and the embodiment does not limit the type of the signal and the signal calculation method. Parameters involved in different signal calculation methods are different, for example, parameters in an adaptive filter algorithm for calculating heart rate are adaptive filter sections, adaptive filter iteration step sizes and the like; parameters in a frequency domain heart rate algorithm for calculating heart rate are frequency search starting frequency, ending frequency, frequency amplitude threshold value and the like, and specific parameter types need to be obtained according to a signal calculation method. The range of the values of the parameters in the signal calculation method can be obtained according to mathematical principles, algorithm experience and signal characteristics.
In this embodiment, after determining the kind of the signal, a corresponding data set is established for the signal, taking the signal as a heart rate signal as an example, the established data set of the heart rate signal needs to satisfy the following three conditions:
1. each data in the data set has a uniform data format including a uniform data length, a uniform data gain, a storage format, etc., so as to improve the standardization of heart rate algorithm design and be suitable for parallel processing platforms.
2. The data set needs to have enough data volume to prevent a small amount of data from causing the overfitting and generalization capability of the heart rate algorithm, and simultaneously avoid the high accuracy of the heart rate algorithm in the calculation result of training data and the low accuracy in the calculation of test data and other labeling data.
3. The data in the data set is complete, i.e. the data set can cover application scene data of various heart rate algorithms, specifically, the data set comprises data of different heart rate values, and the measurement range [ Heartrate ] of the heart rate measurement algorithm is defined according to the physiological characteristics of human bodies or the product requirements low ，HeartRate high ]The true heart rate of the data in the dataset can cover the measurement range [ Heartrate low ，HeartRate high ]Each heart rate value within; secondly, the data set comprises data with different signal quality, physiological characteristics of different people are different, and the generated heart rate signals have different quality, so that the data set needs to contain data with better signal quality and data with poorer signal quality; thirdly, the heart rate algorithm not only can measure the heart rate in a static state, but also can measure the heart rate in a motion state (walking, going upstairs and downstairs, running, climbing, and the like), so that the data set needs to comprise data in the static state, data in different motion states, data in different motion state interferences, data in different measurement positions, and the like.
A specific flow of the method for setting parameters in the signal calculation method of the present embodiment is shown in fig. 1.
Specifically, a parameter space S capable of covering all the parameter values is established, and the parameter space S includes at least one group of parameters whose values are within the parameter value range in the signal calculation method.
And 102, establishing a measurement model for evaluating the quality of the parameters.
Specifically, a measurement model based on the accuracy of the signal calculation method is established to evaluate the merits of the parameters, i.e. for each group of parameters, the measurement model can calculate a measurement value to measure the merits of the group of parameters. The measurement model can be a cost function based on minimum mean square error or a cost function based on least square, wherein the calculated value of the cost function is called a measurement value, and the smaller the measurement value is, the more accurate the signal obtained by calculation is, namely the higher the accuracy of the parameter is, and the closer the accuracy is to a theoretical optimal value; otherwise, the less accurate the calculated signal, i.e. the lower the accuracy of the parameter, the greater the deviation from the theoretical optimum.
Taking a signal as a heart rate signal as an example, a cost function (namely a measurement model) based on the heart rate minimum mean square error is established, and the specific expression of the function is as follows:
wherein J (S) represents the mean square error of the heart rate calculated by the heart rate algorithm when the set parameters are a set of parameters in the parameter space S; HR (HR) s (i, t) represents the heart rate at time t calculated for the ith signal in the dataset when the set of parameters is a set of parameters in the parameter space S; HR (HR) ref (i, t) represents a reference heart rate of an ith signal in the dataset at time t; m represents a total of M signals in the dataset; t denotes the time length T of each signal in the data set.
The process of calculating metric values in a cost function (i.e., metric model) based on heart rate minimum mean square error can be divided into two steps: firstly, configuring a heart rate algorithm according to parameters, and inputting heart rate signals in a data set into the heart rate algorithm to calculate to obtain heart rate values; and substituting the calculated heart rate value and a reference heart rate value known by heart rate signals in the data set into a cost function, wherein the calculated value of the cost function is the metric value.
In the present embodiment, the cost function based on least square error is used as the metric model, but the present embodiment is not limited thereto, and the cost function based on least square may be used as the metric model.
And 103, substituting the preset signals in the signal data set and parameters in the parameter space into a measurement model, and calculating to obtain the parameter quality measurement value.
Specifically, according to the metric model established in step 102, parameters in the signal and parameter space in the preset signal data set are substituted into the metric model, so that the metric value can be calculated and used as the parameter quality metric value. For example, if the metric model is a cost function based on the heart rate minimum mean square error, then the parameter goodness metric values for a set of parameters in the parameter space are: when the set of parameters is set, the heart rate algorithm calculates the mean square error of the heart rate.
And 104, setting parameters in the signal calculation method according to the parameter quality measurement value.
Specifically, a group of parameters with better parameters are set as parameters in the signal calculation method according to the parameter quality measurement values, or parameters corresponding to the parameter quality measurement values satisfying the preset conditions are set as parameters in the signal calculation method, which is not limited in this embodiment.
It should be noted that, in fig. 1, the execution sequence of the steps is only schematically shown, but the actual execution sequence of the steps 101 and 102 is not limited in this embodiment, that is, in this embodiment, the step 102 may be executed first to build a metric model for evaluating the quality of the parameters, and then the step 101 is executed to build a parameter space.
Compared with the prior art, the embodiment establishes a parameter space and establishes a measurement model for evaluating the quality of the parameters, so that the quality degree of the parameters in the parameter space can be evaluated through the measurement model, more accurate parameters are objectively set for a signal calculation method, and subjectivity and uncertainty of manually setting the parameters are avoided.
The second embodiment of the present application relates to a method for setting parameters in a signal calculation method, and the present embodiment refines the first embodiment, mainly in that: a specific implementation of setting parameters in a signal calculation method is provided.
A specific flow of the method for setting parameters in the signal calculation method of the present embodiment is shown in fig. 2.
In step 201, a parameter space is created.
Specifically, the parameter space includes multiple sets of parameters, and the parameter space includes a limited set of parameters, such as N parameters [ p ] in the signal calculation method 1 ，p 2 ，···，p i ，···，p N ]The value range of each parameter is known, the parameter p i The value range of (2) is [ p ] i 1 ，p i 2 ，···，p i K ，···，p i Ki ]Co-K i The possible values are taken.
Further, the parameter space S is defined as a KxN matrix, where N represents the number of parameters,
K i representing parameter p i The number of possible values. Each row of the parameter space S represents a certain combination of N parameters, the parameter space S traverses any combination of N parameter ranges, there are K combinations, and the parameter space S can cover all parameter values. S= [ S ] 1 ，s 2 ，···，s i ，···，s K ] T ，[·] T Representing matrix transpose, s K The K-th combination of parameters, referred to as the K-th set of parameters, is represented, K being greater than 1. For example, there are 3 parameters in the parameter space S, p 1 、p 2 、p 3 Parameter p 1 The value range of (a) is [ a ] 1 ，a 2 ，a 3 ]Parameter p 2 The value range of (b) is 1 ，b 2 ，b 3 ]Parameter p 3 The value range of (c) is 1 ，c 2 ，c 3 ]The parameter space S comprises 27 combinations s= [ a ] 1 b 1 c 1 ，a 1 b 1 c 2 ，a 1 b 1 c 3 ，a 1 b 2 c 1 ，a 1 b 2 c 2 ，a 1 b 2 c 3 ，a 1 b 3 c 1 ，a 1 b 3 c 2 ，a 1 b 3 c 3 ，a 2 b 1 c 1 ，a 2 b 1 c 2 ，a 2 b 1 c 3 ，a 2 b 2 c 1 ，a 2 b 2 c 2 ，a 2 b 2 c 3 ，a 2 b 3 c 1 ，a 2 b 3 c 2 ，a 2 b 3 c 3 ，a 3 b 1 c 1 ，a 3 b 1 c 2 ，a 3 b 1 c 3 ，a 3 b 2 c 1 ，a 3 b 2 c 2 ，a 3 b 2 c 3 ，a 3 b 3 c 1 ，a 3 b 3 c 2 ，a 3 b 3 c 3 ]。
In this embodiment, a discrete parameter value-taking mode, i.e. parameter p, is provided i And (3) taking values at intervals in the value range to form discrete parameter value points.
Preferably, parameter p i The values are equally spaced in the value range, so that the values of all the areas in the value range can be ensured to be taken, and the condition that the value with the better parameter is not taken in a certain area is avoided; the smaller the value interval is, the closer the value of the parameter is to the theoretical optimal value of the parameter, and even the value of the parameter is the theoretical optimal value of the parameter. For example, parameter p i Is 1.4, p when establishing the parameter space S i The value range of (2) is [1,2 ]]When the interval is 0.5, p i The value of (5) is [1,1.5,2 ]]At this time, the parameter p in the finally set signal calculation method i A value of 1.5; when the interval is 0.2, p i The value of (5) is [1,1.2,1.4,1.6,1.8,2 ]]The value of the parameter a in the finally set signal calculation method is the theoretical optimum value 1.4 (i.e. the theoretical optimum value can be obtained).
Specifically, the step is substantially the same as step 102 in the first embodiment, and will not be described herein.
And 203, substituting the parameters in the signal and parameter space in the preset signal data set into the measurement model, and calculating to obtain the parameter quality measurement values of a plurality of groups of parameters.
Specifically, since the parameter space S contains a plurality of sets of parameters S K For example, if the metric model is the cost function based on the heart rate minimum mean square error in step 102 in the first embodiment, the parameter goodness metric value is J (s K ) Representing that the set parameters are the kth group of parameters s K The heart rate algorithm calculates the mean square error of the heart rate. Multiple groups of parameters S in a parameter space S of signals in a preset signal data set K Substituting the heart rate value into a measurement model, specifically taking the measurement model as a cost function based on mean square error as an example, configuring a heart rate algorithm by using a group of parameters of the parameter quality measurement values to be calculated, respectively inputting a plurality of signals in a data set into the configured heart rate algorithm, calculating to obtain heart rate values, and substituting the calculated heart rate values and known reference heart rate values into the cost function to obtain the parameter quality measurement values of the group of parameters; for each group of parameters, the above calculation method is adopted, i.e. multiple groups of parameters s can be calculated respectively K The parameter quality metric of (2).
Preferably, parameters in a signal and parameter space in a preset signal data set are substituted into a measurement model, and parameter quality measurement values of a plurality of groups of parameters are obtained through parallel calculation, which specifically comprises two aspects: firstly, when calculating the parameter quality metric value of each group of parameters, calculating the cost function values corresponding to a plurality of signals in a data set in parallel, and then calculating the parameter quality metric value of the group of parameters; and secondly, when parameter quality metric values of multiple groups of parameters are calculated, parameter quality metric values of each group of parameters in the multiple groups of parameters are calculated in parallel. The parameter quality metric values of a plurality of groups of parameters are obtained through parallel calculation, so that the calculation speed is improved, and the calculation time is shortened.
In step 204, a set of parameters with the smallest parameter quality metric is set as parameters in the signal calculation method.
Specifically, the smaller the parameter quality metric value is, the closer the parameter value is to the optimal value, so that the accuracy of the parameter value can be ensured by setting a group of parameters with the minimum parameter quality metric value as parameters in a signal calculation method; if the parameter with the smallest parameter quality metric value is more than one group, one group can be optionally set as the parameter in the signal calculation method, which is not limited in this embodiment.
It should be noted that, in the present embodiment, a threshold may be set for the parameter quality metric, and a set of parameters is selected from a plurality of sets of parameters corresponding to the parameter quality metric that is smaller than the threshold, and is set as the parameters in the signal calculation method, however, the present embodiment is not limited thereto.
Compared with the first embodiment, the embodiment provides a specific implementation manner of setting parameters in a signal calculation method; namely, selecting an optimal group of parameters from discrete parameter values and setting the optimal group of parameters as parameters in a signal calculation method.
The third embodiment of the present application relates to a method for setting parameters in a signal calculation method, and the present embodiment refines the first embodiment, mainly in that: another specific implementation of setting parameters in the signal calculation method is provided.
A specific flow of the method for setting parameters in the signal calculation method of the present embodiment is shown in fig. 3.
In step 301, a parameter space is established.
Specifically, the parameters in the parameter space S may take any value within its continuous range of values, and the parameter space S may include an infinite set of parameters. For example, there are 3 parameters in the parameter space S, A, B, C, and the value range of the parameter A is [ a ] 1 ，a 3 ]The value range of the parameter B is [ B ] 1 ，b 3 ]The value range of the parameter C is [ C ] 1 ，c 3 ]Then the parameter A may take [ a ] 1 ，a 3 ]The parameter B may take the value of [ B ] 1 ，b 3 ]The parameter C may take the value of [ C ] 1 ，c 3 ]Any one of the values in (a).
In the present embodimentProvides a continuous parameter value-taking mode, namely a parameter p i And continuously taking values in the range of the values.
In step 302, a metric model for evaluating the quality of the parameters is established.
Specifically, the step is substantially the same as step 102 in the embodiment, and will not be described herein.
Specifically, the measurement model J(s), the parameter merit measurement value satisfying the preset condition is minJ(s), and the specific calculation process is as follows:
1. with any set of parameters s in the parameter space established in step 301 K As an initial value for the gradient descent method.
2. The gradient of J (S) with respect to S, i.e.,
3. the update S is performed so that,
where the step size is updated for alpha.
4. Substituting S ' into J (S), calculating to obtain J (S '), and if |J (S ') -J (S) K ) If the I is smaller than the preset threshold, the parameter merit metric J (S') meets the preset condition, and the iteration is stopped; otherwise, let s=s', repeat the above steps 2 to 4 until obtaining the parameter quality metric value satisfying the preset condition. The smaller the preset threshold is set, the closer the parameter corresponding to the obtained parameter merit metric value satisfying the preset condition is to the theoretical optimum value, and the embodiment does not limit this.
Specifically, a set of parameters corresponding to the parameter merit metric values satisfying the preset condition calculated by the gradient descent method in step 303 is set as parameters in the signal calculation method, that is, the parameter S 'corresponding to J (S') is set as parameters in the signal calculation method.
Compared with the first embodiment, the present embodiment provides another specific implementation manner of setting parameters in the signal calculation method, that is, a group of parameters corresponding to the parameter quality measurement values satisfying the preset condition are set as parameters in the signal calculation method.
A fourth embodiment of the present application relates to a method for setting parameters in a signal calculation method, and the present embodiment is an improvement based on the second embodiment, and the main improvement is that: in this embodiment, before step 204 in the second embodiment, the filtering process is performed on the parameter quality metrics of the plurality of sets of parameters.
The specific flow of the parameter setting method in the signal calculation method of the present embodiment is shown in fig. 4.
Step 401 to step 403 are substantially the same as step 201 to step 203, and step 405 is substantially the same as step 204, and are not described herein, but the main difference is that step 404 is added in this embodiment, which is specifically as follows:
Specifically, after the parameter quality metric values of the plurality of sets of parameters are calculated in step 203 (step 403 in this embodiment), the parameter quality metric values of the plurality of sets of parameters are filtered, and then step 405 is performed, where a set of parameters with the smallest parameter quality metric value is set as a parameter in the signal calculation method. The filtering may be mean smoothing or median filtering, but this embodiment is not limited in any way.
Compared with the second embodiment, the embodiment performs filtering processing on the parameter quality measurement values of multiple groups of parameters, so that errors of the parameter quality measurement values caused by noise or abnormal data can be prevented.
The fifth embodiment of the present application relates to a setting device for parameters in a signal calculation method, where the signal may be a heart rate signal, an oxygen signal, and other biological signals, and the signal calculation method may be an adaptive filtering algorithm for calculating a heart rate, a frequency domain heart rate algorithm for calculating a heart rate, an oxygen calculation algorithm for calculating oxygen, and the like, and the embodiment does not limit the type of the signal and the signal calculation method. Parameters involved in different signal calculation methods are different, for example, parameters in an adaptive filter algorithm for calculating heart rate are adaptive filter sections, adaptive filter iteration step sizes and the like; parameters in a frequency domain heart rate algorithm for calculating heart rate are frequency search start-stop frequency, frequency amplitude threshold value and the like, and specific parameter types need to be obtained according to a signal calculation method. The range of the values of the parameters in the signal calculation method can be obtained according to mathematical principles, algorithm experience and signal characteristics.
In this example, referring to fig. 5, the setting device of the parameters in the signal calculation method includes: a first setup module 1, a second setup module 2, a calculation module 3 and a setting module 4.
The first establishing module 1 is used for establishing a parameter space S; the parameter space S includes at least one set of parameters whose values lie within the range of values of the parameters in the signal calculation method.
The second establishing module 2 is configured to establish a metric model for evaluating the merits of the parameters, that is, for each set of parameters, the metric model can calculate a metric value to measure the merits of the set of parameters; the metric model may be a cost function based on minimum mean square error or a cost function based on least squares, where the metric value is the calculated value of the cost function.
The calculating module 3 is configured to substitute the signal in the preset signal data set and the parameter in the parameter space S into the metric model, and calculate to obtain a parameter quality metric value.
The setting module 4 is used for setting parameters in the signal calculation method according to the parameter quality measurement value.
Since the first embodiment corresponds to the present embodiment, the present embodiment can be implemented in cooperation with the first embodiment. The related technical details mentioned in the first embodiment are still valid in this embodiment, and the technical effects that can be achieved in the first embodiment are also achieved in this embodiment, so that the repetition is reduced, and the description is omitted here. Accordingly, the related-art details mentioned in the present embodiment can also be applied to the first embodiment.
Compared with the prior art, the embodiment establishes a parameter space capable of covering all parameter values, and establishes a measurement model for evaluating the quality of the parameters, so that the quality degree of the parameters in the parameter space can be evaluated through the measurement model, more accurate parameters are objectively set for a signal calculation method, and subjectivity and uncertainty of manually setting the parameters are avoided.
The sixth embodiment of the present application relates to a setting device for parameters in a signal calculation method, and the present embodiment is a refinement of the fifth embodiment, and mainly refines that a specific implementation manner for setting parameters in the signal calculation method is provided.
Fig. 5 is a block diagram of a parameter setting device in the signal calculation method in the present embodiment. Wherein the parameter space S established by the first establishing module 1 comprises a plurality of groups of parameters, and the parameter space S comprises a limited group of parameters, such as N parameters [ p ] in the signal calculation method 1 ，p 2 ，···，p i ，···，p N ]The value range of each parameter is known, the parameter p i The value range of (2) is [ p ] i 1 ，p i 2 ，···，p i K ，···，p i Ki ]Co-K i The possible values are taken.
In this embodiment, a discrete parameter value-taking mode, i.e. parameter p, is provided i And (3) taking values at intervals in the value range to form discrete parameter value points.
Preferably, parameter p i The values are equally spaced in the value range, so that the values of all the areas in the value range can be ensured to be obtained, the situation that the value of the parameter with better value is in a certain area but not obtained is avoided, and the smaller the value interval is, the closer the value of the parameter is to the theoretical optimal value of the parameter, and even the value of the parameter is the theoretical optimal value of the parameter. For example, parameter p i Is 1.4, p when establishing the parameter space S i The value range of (2) is [1,2 ]]When the interval is 0.5, p i The value of (5) is [1,1.5,2 ]]At this time, the value of the parameter a in the finally set signal calculation method is 1.5; when the interval is 0.2, p i The value of (5) is [1,1.2,1.4,1.6,1.8,2 ]]The parameter p in the final set signal calculation method i The value is 1.4 of the theoretical optimum (i.e. the theoretical optimum can be obtained).
The calculation module 3 is configured to substitute parameters in a signal and parameter space S in a preset signal data set into a measurement model, specifically, taking the measurement model as a cost function based on a mean square error as an example, configuring a heart rate algorithm with a group of parameters of the parameter quality measurement values to be calculated, then respectively inputting a plurality of signals in the data set into the configured heart rate algorithm and calculating to obtain heart rate values, and substituting the calculated heart rate values and known reference heart rate values into the cost function to obtain the parameter quality measurement values of the group of parameters; for each group of parameters, the above calculation method is adopted, i.e. multiple groups of parameters s can be calculated respectively K The parameter quality metric of (2).
Preferably, the calculating module 3 is configured to substitute the parameters in the signal and parameter space S in the preset signal data set into the metric model, and calculate in parallel to obtain parameter quality metric values of multiple groups of parameters; specifically, the calculation module 3 may include a plurality of calculation units that calculate the cost function values corresponding to the plurality of signals in the data set in parallel when calculating the parameter goodness metric value of each set of parameters; when the parameter quality metric values of the plurality of groups of parameters are calculated, the plurality of groups of parameters are distributed to a plurality of calculation units for parallel calculation. The parameter quality metric values of a plurality of groups of parameters are obtained through parallel calculation, so that the calculation speed is improved, and the calculation time is shortened.
The setting module 4 is configured to set a set of parameters with the smallest parameter quality metric as parameters in the signal calculation method. The smaller the parameter quality metric value is, the closer the parameter value is to the optimal value, so that the accuracy of the parameter value can be ensured by setting a group of parameters with the minimum parameter quality metric value as parameters in a signal calculation method; if the parameter with the smallest parameter quality metric value is more than one group, one group can be optionally set as the parameter in the signal calculation method, which is not limited in this embodiment.
It should be noted that, in the present embodiment, a threshold may be set for the parameter quality metric, and a group of parameters selected from a plurality of groups of parameters corresponding to the parameter quality metric smaller than the threshold may be set as parameters in the signal calculation method, however, the present embodiment is not limited thereto.
Since the second embodiment corresponds to the present embodiment, the present embodiment can be implemented in cooperation with the second embodiment. The related technical details mentioned in the second embodiment are still valid in this embodiment, and the technical effects that can be achieved in the second embodiment are also achieved in this embodiment, so that the repetition is reduced, and the description is omitted here. Accordingly, the related-art details mentioned in the present embodiment can also be applied to the second embodiment.
Compared with the fifth embodiment, the present embodiment provides a specific implementation manner of setting parameters in the signal calculation method; namely, selecting an optimal group of parameters from discrete parameter values and setting the optimal group of parameters as parameters in a signal calculation method.
The seventh embodiment of the present application relates to a setting device for parameters in a signal calculation method, and the present embodiment is a refinement of the fifth embodiment, and mainly refines that another specific implementation manner for setting parameters in the signal calculation method is provided.
Fig. 5 is a block diagram of a parameter setting device in the signal calculation method in the present embodiment. When the first building module 1 builds the parameter space S, the parameters in the parameter space S may take any values in the continuous value range, and the parameter space S may include an infinite set of parameters. For example, there are 3 parameters in the parameter space S, p 1 、p 2 、p 3 Parameter p 1 The value range of (a) is [ a ] 1 ，a 3 ]Parameter p 2 The value range of (b) is 1 ，b 3 ]Parameter p 3 The value range of (c) is 1 ，c 3 ]Parameter p 1 Can take [ a ] 1 ，a 3 ]Any one of the values of parameter p 2 Can take [ b ] 1 ，b 3 ]Any one of the values of parameter p 3 Can take [ c ] 1 ，c 3 ]Any one of the values in (a).
In this embodiment, a continuous parameter value-taking mode, i.e. parameter p, is provided i And continuously taking values in the range of the values.
The calculating module 3 is used for substituting the parameters in the signal and parameter space S in the preset signal data set into the measurement model, and calculating by adopting a gradient descent method to obtain the parameter quality measurement value meeting the preset condition; wherein, any group of parameters in the parameter space S is taken as the initial value of the gradient descent method.
The setting module 4 is configured to set a set of parameters corresponding to the parameter quality metric values satisfying the preset condition as parameters in the signal calculation method.
Since the third embodiment corresponds to the present embodiment, the present embodiment can be implemented in cooperation with the third embodiment. The related technical details mentioned in the third embodiment are still valid in this embodiment, and the technical effects achieved in the third embodiment may also be achieved in this embodiment, so that the repetition is reduced, and the description is omitted here. Accordingly, the related-art details mentioned in the present embodiment can also be applied to the third embodiment.
In this embodiment, compared to the fifth embodiment, another specific implementation manner of setting parameters in the signal calculation method is provided, that is, a set of parameters corresponding to the parameter quality metric values satisfying the preset condition is set as the parameters in the signal calculation method.
An eighth embodiment of the present application relates to a device for setting parameters in a signal calculation method, and the present embodiment is an improvement based on the fifth embodiment, and the main improvement is that: in this embodiment, referring to fig. 6, the setting device of the parameters in the signal calculation method further includes a filtering module 5.
The filtering module 5 is used for filtering the parameter quality measurement values of the plurality of groups of parameters. The filtering may be mean-smoothing or median filtering, however, the present embodiment is not limited in any way.
Since the fourth embodiment corresponds to the present embodiment, the present embodiment can be implemented in cooperation with the fourth embodiment. The related-art details mentioned in the fourth embodiment are still valid in the present embodiment, the technical effects achieved in the fourth embodiment are also achieved in the present embodiment, and in order to reduce repetition, a detailed description is omitted here. Accordingly, the related-art details mentioned in the present embodiment can also be applied to the fourth embodiment.
In this embodiment, compared to the sixth embodiment, the filtering process is performed on the parameter quality metrics of the plurality of sets of parameters, so that errors of the parameter quality metrics due to noise or abnormal data can be prevented.
It will be understood by those of ordinary skill in the art that the foregoing embodiments are specific embodiments in which the present application is implemented and that various changes in form and details may be made therein without departing from the spirit and scope of the present application.
Claims (10)
1. A method for setting parameters in a signal calculation method, comprising:
establishing a parameter space, wherein the parameter space comprises at least one group of parameters with values within a parameter value range in the signal calculation method;
establishing a measurement model for evaluating the quality of the parameters; the measurement model is a cost function based on minimum mean square error or a cost function based on least square;
substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain a parameter quality measurement value;
setting parameters in the signal calculation method according to the parameter quality measurement values;
wherein the parameter space comprises a plurality of groups of parameters; substituting the signal in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain a parameter quality measurement value, wherein the parameter quality measurement value specifically comprises: substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain parameter quality measurement values of a plurality of groups of parameters;
the parameters in the signal calculation method are set according to the parameter quality measurement values, specifically: setting a group of parameters with the smallest parameter quality measurement value as parameters in the signal calculation method;
the calculation mode of the parameter quality measurement value is as follows: configuring the signal calculation method according to parameters in the parameter space, and substituting signals in the preset signal data set into the signal calculation method to obtain signal values; substituting the calculated signal value and the reference signal value in the preset signal data set into the measurement model, and calculating to obtain the parameter quality measurement value.
2. The method of claim 1, wherein the substituting the signal in the preset signal data set and the parameters in the parameter space into the metric model, and calculating to obtain the parameter quality metric values of the plurality of groups of parameters comprises:
substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating in parallel to obtain the parameter quality measurement values of a plurality of groups of parameters.
3. The method of claim 1, wherein before setting the set of parameters with the smallest parameter goodness measure as the parameters in the signal calculation method, further comprising:
and carrying out filtering treatment on the parameter quality measurement values of the plurality of groups of parameters.
4. A method according to claim 3, wherein the filtering is mean smoothing or median filtering.
5. A method for setting parameters in a signal calculation method, comprising:
establishing a parameter space, wherein the parameter space comprises at least one group of parameters with values within a parameter value range in the signal calculation method;
establishing a measurement model for evaluating the quality of the parameters; the measurement model is a cost function based on minimum mean square error or a cost function based on least square;
substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain a parameter quality measurement value;
setting parameters in the signal calculation method according to the parameter quality measurement values;
substituting the signal in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain a parameter quality measurement value, wherein the parameter quality measurement value is specifically: substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating by adopting a gradient descent method to obtain a parameter quality measurement value meeting preset conditions; wherein, any group of parameters in the parameter space is used as the initial value of the gradient descent method;
the parameters in the signal calculation method are set according to the parameter quality measurement values, specifically: setting a group of parameters corresponding to the parameter quality measurement values meeting preset conditions as parameters in the signal calculation method;
the calculation mode of the parameter quality measurement value is as follows: configuring the signal calculation method according to parameters in the parameter space, and substituting signals in the preset signal data set into the signal calculation method to obtain signal values; substituting the calculated signal value and the reference signal value in the preset signal data set into the measurement model, and calculating to obtain the parameter quality measurement value.
6. A setting device of parameters in a signal calculation method, comprising:
the first establishing module is used for establishing a parameter space; the parameter space comprises at least one group of parameters with values within the parameter value range in the signal calculation method;
the second building module is used for building a measurement model for evaluating the quality of the parameters; the measurement model is a cost function based on minimum mean square error or a cost function based on least square;
the calculation module is used for substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain the parameter quality measurement value;
the setting module is used for setting parameters in the signal calculation method according to the parameter quality measurement value;
the parameter space comprises a plurality of groups of parameters; the calculation module is used for substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain parameter quality measurement values of multiple groups of parameters; the setting module is used for setting a group of parameters with the smallest parameter quality measurement value as parameters in the signal calculation method;
the calculation module is further configured to configure the signal calculation method according to parameters in the parameter space, and substitutes signals in the preset signal data set into the signal calculation method to obtain signal values; substituting the calculated signal value and the reference signal value in the preset signal data set into the measurement model, and calculating to obtain the parameter quality measurement value.
7. The apparatus of claim 6, wherein the computing module is configured to substitute the signal in the preset signal data set and the parameters in the parameter space into the metric model, and calculate in parallel to obtain the parameter quality metric values of the plurality of groups of parameters.
8. The apparatus of claim 6, wherein the apparatus further comprises a filtering module; the filtering module is used for filtering the parameter quality measurement values of the plurality of groups of parameters.
9. The apparatus of claim 8, wherein the filtering is mean-smoothing or median filtering.
10. A setting device of parameters in a signal calculation method, comprising:
the first establishing module is used for establishing a parameter space; the parameter space comprises at least one group of parameters with values within the parameter value range in the signal calculation method;
the second building module is used for building a measurement model for evaluating the quality of the parameters; the measurement model is a cost function based on minimum mean square error or a cost function based on least square;
the calculation module is used for substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating to obtain the parameter quality measurement value;
the setting module is used for setting parameters in the signal calculation method according to the parameter quality measurement value;
the calculation module is used for substituting the signals in the preset signal data set and the parameters in the parameter space into the measurement model, and calculating by adopting a gradient descent method to obtain the parameter quality measurement value meeting the preset condition; wherein, any group of parameters in the parameter space is used as the initial value of the gradient descent method; the setting module is used for setting a group of parameters corresponding to the parameter quality measurement values meeting preset conditions as parameters in the signal calculation method;
the calculation module is further configured to configure the signal calculation method according to parameters in the parameter space, and substitutes signals in the preset signal data set into the signal calculation method to obtain signal values; substituting the calculated signal value and the reference signal value in the preset signal data set into the measurement model, and calculating to obtain the parameter quality measurement value.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2017/093909 WO2019014933A1 (en) | 2017-07-21 | 2017-07-21 | Method and device for setting parameters in signal calculation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109791564A CN109791564A (en) | 2019-05-21 |
| CN109791564B true CN109791564B (en) | 2023-06-16 |
Family
ID=65014839
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201780000780.1A Active CN109791564B (en) | 2017-07-21 | 2017-07-21 | Method and device for setting parameters in signal calculation method |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109791564B (en) |
| WO (1) | WO2019014933A1 (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010067830A (en) * | 2008-09-11 | 2010-03-25 | Fujitsu Ltd | Model parameter abstraction program |
| CN102663417A (en) * | 2012-03-19 | 2012-09-12 | 河南工业大学 | Feature selection method for pattern recognition of small sample data |
| CN104995562A (en) * | 2012-12-18 | 2015-10-21 | 科磊股份有限公司 | Integrated use of model-based metrology and a process model |
| CN105335619A (en) * | 2015-11-10 | 2016-02-17 | 广西大学 | Collaborative optimization method applicable to parameter back analysis of high calculation cost numerical calculation model |
| WO2017016086A1 (en) * | 2015-07-30 | 2017-02-02 | 华南理工大学 | Depression evaluating system and method based on physiological information |
| CN106599230A (en) * | 2016-12-19 | 2017-04-26 | 北京天元创新科技有限公司 | Method and system for evaluating distributed data mining model |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6456675B2 (en) * | 1999-10-12 | 2002-09-24 | Memorylink Corporation | Diversity reception employing periodic testing |
| US20090171226A1 (en) * | 2007-12-31 | 2009-07-02 | Nellcor Puritan Bennett Llc | System and method for evaluating variation in the timing of physiological events |
| CN102722103A (en) * | 2012-05-18 | 2012-10-10 | 镇江雅迅软件有限责任公司 | Method for optimizing multi-objective parameters based on fuzzy analysis and visualization technology |
| US20140301309A1 (en) * | 2013-04-08 | 2014-10-09 | Qualcomm Incorporated | Low complexity blind detection of transmission parameters of interferers |
| CN104200087B (en) * | 2014-06-05 | 2018-10-02 | 清华大学 | For the parameter optimization of machine learning and the method and system of feature tuning |
-
2017
- 2017-07-21 CN CN201780000780.1A patent/CN109791564B/en active Active
- 2017-07-21 WO PCT/CN2017/093909 patent/WO2019014933A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010067830A (en) * | 2008-09-11 | 2010-03-25 | Fujitsu Ltd | Model parameter abstraction program |
| CN102663417A (en) * | 2012-03-19 | 2012-09-12 | 河南工业大学 | Feature selection method for pattern recognition of small sample data |
| CN104995562A (en) * | 2012-12-18 | 2015-10-21 | 科磊股份有限公司 | Integrated use of model-based metrology and a process model |
| WO2017016086A1 (en) * | 2015-07-30 | 2017-02-02 | 华南理工大学 | Depression evaluating system and method based on physiological information |
| CN105335619A (en) * | 2015-11-10 | 2016-02-17 | 广西大学 | Collaborative optimization method applicable to parameter back analysis of high calculation cost numerical calculation model |
| CN106599230A (en) * | 2016-12-19 | 2017-04-26 | 北京天元创新科技有限公司 | Method and system for evaluating distributed data mining model |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019014933A1 (en) | 2019-01-24 |
| CN109791564A (en) | 2019-05-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107949321B (en) | Temporal interference removal and improved heart rate measurement tracking mechanism | |
| AU720450B2 (en) | Method and apparatus for adaptively averaging data signals | |
| JP6971150B2 (en) | Methods and devices for deriving the mean arterial pressure of a subject | |
| US9936886B2 (en) | Method for the estimation of the heart-rate and corresponding system | |
| CN108478206B (en) | Heart rate monitoring method based on pulse wave in motion state | |
| ATE341993T1 (en) | DEVICE FOR DETERMINING LEFT VENTRICULAR EJECTION TIME IN A HEART | |
| Yang et al. | Estimation and validation of arterial blood pressure using photoplethysmogram morphology features in conjunction with pulse arrival time in large open databases | |
| US10744261B2 (en) | System and method of a biosensor for detection of vasodilation | |
| CN108261190A (en) | Continuous BP measurement method, apparatus and equipment | |
| CN109497977A (en) | Human heart rate and method for detecting blood oxygen saturation and device | |
| US10470717B2 (en) | Pulse validation | |
| WO2020114190A1 (en) | Recursion method and system for confidence degree of blood oxygen saturation measurement, and storage medium | |
| CN116269269A (en) | Continuous blood pressure measuring instrument, system and method with data quality evaluation function | |
| JP7247444B2 (en) | device for measuring blood pressure | |
| CN111466899A (en) | Continuous blood pressure non-invasive monitoring method based on MPGA-MIV-BP model cluster | |
| CN109791564B (en) | Method and device for setting parameters in signal calculation method | |
| US20200065649A1 (en) | Method for robust and noise-tolerant SpO2 determination | |
| JP2017023472A (en) | Bio-information estimation device, bio-information estimation method, and computer program | |
| Everson et al. | BioTranslator: inferring R-peaks from ambulatory wrist-worn PPG signal | |
| Gircys et al. | Movement artefact resistant photoplethysmographic probe | |
| CN107223037A (en) | Wearable device, the method and device for eliminating motion artifacts | |
| CN109643579B (en) | Quality evaluation method and device, model building method and module and wearable device | |
| CN111000544A (en) | Method and system for constructing hybrid continuous blood pressure measurement model based on PPG waveform | |
| Zhang et al. | Non-invasive blood glucose detection using NIR based on GA and SVR | |
| Fedorin et al. | Lightweight neural network based model for real-time precise HR monitoring during high intensity workout using consumer smartwatches |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |