CN105796091B - A kind of intelligent terminal for removing electrocardiosignal vehicle movement noise - Google Patents
A kind of intelligent terminal for removing electrocardiosignal vehicle movement noise Download PDFInfo
- Publication number
- CN105796091B CN105796091B CN201610113597.XA CN201610113597A CN105796091B CN 105796091 B CN105796091 B CN 105796091B CN 201610113597 A CN201610113597 A CN 201610113597A CN 105796091 B CN105796091 B CN 105796091B
- Authority
- CN
- China
- Prior art keywords
- electrocardiosignal
- channel
- neural network
- electrocardio
- segment
- 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.)
- Expired - Fee Related
Links
- 238000012549 training Methods 0.000 claims abstract description 58
- 230000001133 acceleration Effects 0.000 claims abstract description 57
- 238000013528 artificial neural network Methods 0.000 claims abstract description 55
- 230000003068 static effect Effects 0.000 claims abstract description 15
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000004891 communication Methods 0.000 claims abstract description 9
- 238000006243 chemical reaction Methods 0.000 claims abstract description 7
- 238000000034 method Methods 0.000 claims abstract description 5
- 238000012360 testing method Methods 0.000 claims description 17
- 238000001514 detection method Methods 0.000 claims description 14
- 238000004088 simulation Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 10
- 210000002569 neuron Anatomy 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000002939 conjugate gradient method Methods 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 12
- 238000003745 diagnosis Methods 0.000 abstract description 4
- 230000001537 neural effect Effects 0.000 abstract 2
- 208000024172 Cardiovascular disease Diseases 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 108010076504 Protein Sorting Signals Proteins 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012806 monitoring device Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 206010003119 arrhythmia Diseases 0.000 description 1
- 230000006793 arrhythmia Effects 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 235000012054 meals Nutrition 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 208000031225 myocardial ischemia Diseases 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
- A61B5/316—Modalities, i.e. specific diagnostic methods
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
- A61B5/7207—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal of noise induced by motion artifacts
- A61B5/721—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal of noise induced by motion artifacts using a separate sensor to detect motion or using motion information derived from signals other than the physiological signal to be measured
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Pathology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Cardiology (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a kind of intelligent terminals for removing electrocardiosignal vehicle movement noise, including processor and bluetooth communication unit；Processor is loaded with signal acquisition module, preprocessing module, neural metwork training module and denoising module.Intelligent terminal of the present invention acquires exercise data as neural network input sample using 3-axis acceleration sensor, it is pre-processed by the dynamic electrocardiogram (ECG) data of format conversion and method for normalizing before neural metwork training when and movement static to human body and the deviation that makes between the two is used as the supervision of neural network, and then noise model is established with the optimal way of ratio conjugate gradient using artificial neural network learning algorithm, the dynamic electrocardiogram (ECG) data obtained with monitoring subtracts corresponding motion artifacts and completes denoising.Whole equipment is easy and is easily achieved, and vehicle movement noise can be effectively automatically removed during cardioelectric monitor, obtains accurate electrocardiogram (ECG) data, guarantee is provided to the diagnosis and treatment of angiocardiopathy for doctor.
Description
Technical Field
The invention belongs to the technical field of medical equipment, and particularly relates to an intelligent terminal for removing electrocardiosignal vehicle motion noise.
Background
The dynamic electrocardiogram is used for continuously recording 24-hour dynamic electrocardiogram data, including electrocardiogram data under different conditions of rest, activity, work, meal and the like, can help to find symptoms such as arrhythmia, myocardial ischemia and the like, and provides important basis for clinical analysis of illness state and diagnosis and treatment. Along with the emergence of portable wearable electrocardio monitoring equipment, the user can carry out very convenient electrocardio monitoring under various states. However, the ecg signals are easily interfered by various noises, such as ac interference, emg interference, baseline wander, wrong connection of lead wires, insufficient contact between electrode plates and skin, etc. with the development of technology, most ecg devices are equipped with various filters to eliminate these interferences, but when the portable ecg devices are used in mobile vehicles, the interference of vehicle movement may severely limit the quality of the ecg signals.
The motion interference of the vehicle can act on objects on the vehicle, including electrocardio monitoring equipment worn by a human body. The motion state of the vehicle can well reflect the influence of interference on the electrocardio monitoring equipment, and the motion state of the vehicle can be very conveniently acquired through the triaxial acceleration sensor, so that the influence mode and degree of motion are analyzed, and a data basis is provided for eliminating the motion interference of the vehicle.
The triaxial acceleration sensor is a sensor for converting physical signal acceleration into an electric signal convenient for measurement, the working principle of most of the existing triaxial acceleration sensors is piezoresistive type, piezoelectric type and capacitance type, the generated acceleration is in direct proportion to the change of resistance, voltage and capacitance, and the value of the acceleration can be calculated by calculating the relation between the change and the acceleration. The three-axis acceleration sensor can acquire coordinate components of the three-axis acceleration sensor by detecting signals of x, y and z axes on the occasion that the motion state of an object is not known in advance, so that the motion state of the object is accurately judged, the three-axis acceleration sensor has the characteristic points of small volume and light weight, can measure the space acceleration, can comprehensively and accurately reflect the motion property of the object, and is widely applied to the fields of aerospace, automobiles, robots, medicine and the like.
Most portable dynamic electrocardiogram monitoring devices in the market at present have simple structures, do not have the function of eliminating vehicle motion noise, are easy to interfere when used in a vehicle motion state to cause instability of dynamic electrocardiogram signals, and recorded electrocardiogram waveforms cannot be used for electrocardiogram analysis, so that the application range of the dynamic electrocardiogram monitoring devices is limited, and doctors can diagnose cardiovascular disease patients to cause serious consequences.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides the intelligent terminal for removing the electrocardiosignal vehicle motion noise, which can effectively and automatically remove the vehicle motion noise in the electrocardio monitoring process, obtain accurate electrocardio data and provide guarantee for the diagnosis and treatment of cardiovascular diseases by doctors.
The intelligent terminal for removing electrocardiosignal vehicle motion noise comprises a processor and a Bluetooth communication unit, wherein the processor is loaded with the following functional modules:
the signal acquisition module is used for collecting electrocardio data provided by the electrocardio monitor, training data corresponding to each heart rate state provided by the electrocardio signal simulation generator and motion data provided by the triaxial acceleration sensor through a Bluetooth communication module in the intelligent terminal; the electrocardio data is a three-channel electrocardio detection signal segment which is obtained by the user through the detection of an electrocardio monitor when the vehicle moves in daily life; the training data comprises m groups of three-channel static electrocardiosignal segments which are generated by the electrocardiosignal simulation generator when the electrocardiosignal simulation generator is static and are in any heart rate state and m groups of three-channel moving electrocardiosignal segments which are generated by the electrocardiosignal simulation generator when the vehicle moves and are in any heart rate state, wherein m is a natural number more than 1; the motion data comprises a triaxial acceleration sensor which synchronously acquires a triaxial acceleration signal segment corresponding to a three-channel motion electrocardiosignal segment generated by the electrocardiosignal simulation generator when the vehicle moves and a triaxial acceleration signal segment corresponding to a three-channel electrocardiosignal detection signal segment detected by the electrocardiosignal monitor when the vehicle moves;
the preprocessing module is used for preprocessing the three-channel static electrocardiosignal segment, the three-channel moving electrocardiosignal segment, the three-channel electrocardio detection signal segment and the three-axis acceleration signal segment; meanwhile, for any heart rate state, correspondingly subtracting the three-channel moving electrocardiosignal segment preprocessed in the heart rate state from the three-channel static electrocardiosignal segment to obtain m groups of three-channel electrocardio-moving noises;
the neural network training module is used for training any heart rate state according to three-channel electrocardio-motion noise in the heart rate state and the preprocessed triaxial acceleration signal segment through an artificial neural network learning algorithm to obtain a vehicle motion noise model related to the three-lead electrocardio-signals in the heart rate state;
and the denoising module is used for inputting the triaxial acceleration signal segment corresponding to the three-channel electrocardio detection signal segment into the corresponding vehicle motion noise model according to the heart rate state of the user to output to obtain a group of three-channel electrocardio motion noises, and further subtracting the three-channel electrocardio motion noises from the three-channel electrocardio detection signal segment to obtain three-channel electrocardio data with the vehicle motion noises eliminated.
The preprocessing module carries out preprocessing including format conversion and normalization processing on the three-channel stationary electrocardiosignal section, the three-channel moving electrocardiosignal section, the three-channel electrocardio detection signal section and the three-axis acceleration signal section so as to obtain data signals in proper formats and ranges.
The preprocessing module performs normalization processing based on the following expression:
wherein: x is any signal value in the signal segment, xminAnd xmaxThe minimum value and the maximum value in the signal section are respectively, and x' is a signal value after normalization processing of the corresponding x.
The artificial neural network learning algorithm adopted by the neural network training module takes a proportional conjugate gradient method as an optimization direction.
The specific process of training by the neural network training module through an artificial neural network learning algorithm is as follows:
(1) dividing m groups of three-axis acceleration signal segments corresponding to the three-channel sports electrocardiosignal segments after pretreatment into a training set and a test set, wherein the training set is larger than the test set;
(2) initializing and constructing a neural network consisting of an input layer, a hidden layer and an output layer;
(3) a group of triaxial acceleration signal segments are arbitrarily selected from the training set and substituted into the neural network to calculate and obtain an output result corresponding to the electrocardio-motion noise, and the accumulated error between the output result and the three-channel electrocardio-motion noise corresponding to the triaxial acceleration signal segments is calculated;
(4) correcting the weights between an input layer and a hidden layer and between the hidden layer and an output layer in the neural network through a gradient descent method according to the accumulated error, and then arbitrarily taking a next group of triaxial acceleration signal segments from the training set to substitute the corrected neural network;
(5) and (4) traversing all the triaxial acceleration signal segments in the training set according to the steps (3) and (4), and taking the neural network corresponding to the minimum accumulated error as a vehicle motion noise model.
The input layer of the neural network constructed by the neural network training module through initialization consists of 3 neurons, the hidden layer consists of 10 neurons, and the output layer consists of 3 neurons.
The expression of the neuron function g (z) in the neural network initially constructed by the neural network training module is as follows:
wherein: z is the argument of the function.
The neural network training module substitutes the triaxial acceleration signal segments in the test set into the vehicle motion noise model one by one to obtain an output result corresponding to the electrocardio-motion noise for the vehicle motion noise model obtained through training, so that the output result corresponding to each group of triaxial acceleration signal segments in the test set is compared with the three-channel electrocardio-motion noise, and if the comparison result of a certain proportion of triaxial acceleration signal segments in the test set is less than or equal to a threshold value, the vehicle motion noise model is finally determined; otherwise, the electrocardiosignal simulation generator and the three-axis acceleration sensor are used for acquiring more three-channel static electrocardiosignal segments, three-channel moving electrocardiosignal segments and three-axis acceleration signal segments, and the three-channel moving electrocardiosignal segments and the three-axis acceleration signal segments are used for training as the input of the neural network training module by increasing the scale of the training set after passing through the signal acquisition module and the preprocessing module in the intelligent terminal.
The intelligent terminal can be a smart phone, a tablet computer or a PC.
The intelligent terminal of the invention utilizes a triaxial acceleration sensor to collect motion data as a neural network input sample, preprocesses dynamic electrocardio data of a human body when the human body is static and moving before the neural network training through a format conversion and normalization method, leads the deviation between the dynamic electrocardio data and the neural network to be used as the supervision of the neural network, further utilizes an artificial neural network learning algorithm to establish a noise model in an optimization mode of proportional conjugate gradient, and subtracts corresponding motion noise from the dynamic electrocardio data obtained by monitoring to finish denoising. The whole equipment is simple and easy to realize, can effectively and automatically remove vehicle motion noise in the electrocardio monitoring process, obtains accurate electrocardio data, and provides guarantee for the diagnosis and treatment of cardiovascular diseases by doctors.
Drawings
Fig. 1 is a schematic structural diagram of an intelligent terminal according to the present invention.
FIG. 2 is a schematic diagram of an artificial neural network model in the electrocardiosignal denoising process according to the present invention.
Detailed Description
In order to more specifically describe the present invention, the following detailed description is provided for the technical solution of the present invention with reference to the accompanying drawings and the specific embodiments.
As shown in fig. 1, in this embodiment, the intelligent terminal for removing the electrocardiosignal vehicle motion noise adopts a smart phone, and the smart phone includes a processor and a bluetooth communication module, and the bluetooth communication module is connected to the processor; the processor comprises a signal acquisition module, a preprocessing module, a neural network training module and a denoising module; wherein:
the Bluetooth communication module is used for communicating the intelligent terminal with the electrocardio monitoring terminal, the intelligent terminal sends a command to the monitoring terminal, and the monitoring terminal responds to the command and uploads electrocardio data to the intelligent terminal.
The signal acquisition module is used for receiving dynamic electrocardiogram data from the generator or the monitor and motion state data from the acceleration sensor through the Bluetooth communication module; the dynamic electrocardiogram data comprises m groups of electrocardiogram signal sequences E which are acquired in a static state and are in 1 heart rate state, and n × m groups of static electrocardiogram data E exist due to n different heart rate states; the M groups of electrocardiosignal sequences M which are collected in the vehicle motion state and are in 1 heart rate state have n × M groups of motion electrocardio data M due to n different heart rate states.
The motion state data are acquired by m groups of motion states of the vehicle and correspond to output signal sequences Q [ x, y, z ] of the three-axis acceleration sensor under 1 heart rate state, and as n different heart rate states exist, n-m groups of motion state sequences Q [ x, y, z ] are provided, and m and n are natural numbers larger than 1.
The preprocessing module is used for carrying out format conversion and normalization processing on the dynamic electrocardiogram data and the motion state data to obtain original data with proper format and range. In this embodiment, the data sampling rate is 250, the AD conversion bit number is 24 bits, the sampling rate is reduced to 200 by the down-sampling algorithm, and the 24-bit data is converted to 16 bits by the data compression algorithm, so as to obtain data with smaller capacity, but the requirement of the neural network training module needs to be met. The normalization algorithm adopts a linear conversion algorithm, and the expression is as follows:
wherein: x is the input vector, max is the maximum value of x, min is the minimum value of x, and f (x) is the output vector after normalization.
The neural network training module trains and tests the n x m groups of preprocessed motion state working data sequences through an artificial neural network deep learning algorithm to obtain a noise model; the specific implementation mode is as follows:
(1) and dividing m groups of preprocessed motion state data sequences Q [ x, y, z ] in 1 heart rate state into a training set and a test set, wherein the training set is larger than the test set.
(2) Establishing an initial neural network model according to an artificial neural network learning algorithm: the neural network model is composed of an input layer, a hidden layer and an output layer, wherein the input layer is connected with the hidden layer through a formula (1), the input of the input layer is a working data sequence of x, y and z axes of a three-axis acceleration sensor, an activation function of the hidden layer and the output layer is a formula (2), the hidden layer is composed of 10 neurons, the output of the output layer is noise of the x, y and z axes of the three-axis acceleration sensor, and the built neural network model is shown in figure 2.
Wherein,andfor the coefficients of the hidden layer, g is the tansig function, i is 1,2, …,10, j is 1,2, 3.
Wherein,andthe coefficients of the output layers are i 1,2,3, and j 1,2, …, 10.
(3) And inputting a group of sample data of the training set into the neural network under the current weight coefficient, and sequentially calculating the output of each node of the input layer, the hidden layer and the output layer.
(4) Calculating the accumulated error E between the output layers of all the training samples and the expected results of the training samples according to a formula (3), correcting the weight coefficient between each node of the input layer and the hidden layer according to a formula (4) and a conjugate gradient descent method according to a formula (5):
wherein: e is the accumulated error, and E is the accumulated error,for the kth output, o, of a single training sample through the output layer of the neural networkkFor the kth expected result of a single training sample, m is the total number of training set samples, and p is the total number of output layers.
Wherein: w is aho(t) is the weight coefficient between the hidden layer and the output layer when the tth sample is input into the neural network,for the output of a single training sample through the output layer of the neural network, o is the expected result of the single training sample, xhα is the learning rate for the output of the hidden layer.
Wherein: w is aih(t) is the weight coefficient between the input layer and the hidden layer when the t-th sample is input into the neural network, xiIs the output of the input layer.
(5) Repeating the step (3) and the step (4) to traverse all the training sets, establishing a noise model, obtaining the noise model corresponding to the weight coefficient group when the E is minimum, then testing the noise model by using the testing set, and if the testing accuracy is higher than a threshold value, determining the noise model as the optimal model; and if the accuracy of the test is lower than the threshold value, continuing to increase the training samples, and repeating the step (3) and the step (4) to train the neural network until the training is completed. The weight coefficient matrix obtained by training in this embodiment is:
input layer and hidden layer:
hidden layer and output layer:
the denoising module is used for restoring a system function of the noise model according to the weight proportion of each layer of the neural network; and substituting triaxial acceleration output data which are obtained by daily monitoring of a user and reflect the motion state of the human body into the noise model, outputting a corresponding noise result, and subtracting a noise value output by the noise model from the dynamic electrocardiogram data obtained by the monitor to obtain the dynamic electrocardiogram data with noise removed.
The embodiments described above are intended to facilitate one of ordinary skill in the art in understanding and using the invention. It will be readily apparent to those skilled in the art that various modifications to the above-described embodiments may be made, and the generic principles defined herein may be applied to other embodiments without the use of inventive faculty. Therefore, the present invention is not limited to the above embodiments, and those skilled in the art should make improvements and modifications to the present invention based on the disclosure of the present invention within the protection scope of the present invention.
Claims (5)
1. An intelligent terminal for removing electrocardiosignal vehicle motion noise comprises a processor and a Bluetooth communication unit; the processor is characterized in that the processor is loaded with the following functional modules:
the signal acquisition module is used for collecting electrocardio data provided by the electrocardio monitor, training data corresponding to each heart rate state provided by the electrocardio signal simulation generator and motion data provided by the triaxial acceleration sensor through a Bluetooth communication module in the intelligent terminal; the electrocardio data is a three-channel electrocardio detection signal segment which is obtained by the user through the detection of an electrocardio monitor when the vehicle moves in daily life; the training data comprises m groups of three-channel static electrocardiosignal segments which are generated by the electrocardiosignal simulation generator when the electrocardiosignal simulation generator is static and are in any heart rate state and m groups of three-channel moving electrocardiosignal segments which are generated by the electrocardiosignal simulation generator when the vehicle moves and are in any heart rate state, wherein m is a natural number more than 1; the motion data comprises a triaxial acceleration sensor which synchronously acquires a triaxial acceleration signal segment corresponding to a three-channel motion electrocardiosignal segment generated by the electrocardiosignal simulation generator when the vehicle moves and a triaxial acceleration signal segment corresponding to a three-channel electrocardiosignal detection signal segment detected by the electrocardiosignal monitor when the vehicle moves;
the preprocessing module is used for preprocessing the three-channel static electrocardiosignal segment, the three-channel moving electrocardiosignal segment, the three-channel electrocardio detection signal segment and the three-axis acceleration signal segment; meanwhile, for any heart rate state, correspondingly subtracting the three-channel moving electrocardiosignal segment preprocessed in the heart rate state from the three-channel static electrocardiosignal segment to obtain m groups of three-channel electrocardio-moving noises;
the neural network training module is used for training any heart rate state according to three-channel electrocardio-motion noise in the heart rate state and the preprocessed triaxial acceleration signal segment through an artificial neural network learning algorithm to obtain a vehicle motion noise model related to the three-lead electrocardio-signals in the heart rate state;
the de-noising module is used for inputting the triaxial acceleration signal segment corresponding to the three-channel electrocardio detection signal segment into the corresponding vehicle motion noise model according to the heart rate state of the user and outputting the triaxial acceleration signal segment to obtain a group of three-channel electrocardio motion noise, and then the three-channel electrocardio motion noise is subtracted from the three-channel electrocardio detection signal segment to obtain three-channel electrocardio data with the vehicle motion noise eliminated;
the preprocessing module is used for preprocessing a three-channel static electrocardiosignal segment, a three-channel moving electrocardiosignal segment, a three-channel electrocardio detection signal segment and a three-axis acceleration signal segment, including format conversion and normalization processing, so as to obtain data signals in proper formats and ranges; wherein the normalization process is based on the following expression:
wherein: x is any signal value in the signal segment, xminAnd xmaxRespectively the minimum value and the maximum value in the signal section, wherein x' is the signal value of the corresponding x after normalization processing;
the specific process of training by the neural network training module through an artificial neural network learning algorithm is as follows:
(1) dividing m groups of three-axis acceleration signal segments corresponding to the three-channel sports electrocardiosignal segments after pretreatment into a training set and a test set, wherein the training set is larger than the test set;
(2) initializing and constructing a neural network consisting of an input layer, a hidden layer and an output layer;
(3) a group of triaxial acceleration signal segments are arbitrarily selected from the training set and substituted into the neural network to calculate and obtain an output result corresponding to the electrocardio-motion noise, and the accumulated error between the output result and the three-channel electrocardio-motion noise corresponding to the triaxial acceleration signal segments is calculated;
(4) correcting the weights between an input layer and a hidden layer and between the hidden layer and an output layer in the neural network through a gradient descent method according to the accumulated error, and then arbitrarily taking a next group of triaxial acceleration signal segments from the training set to substitute the corrected neural network;
(5) traversing all the triaxial acceleration signal segments in the training set according to the steps (3) and (4), and taking a neural network corresponding to the smallest accumulated error as a vehicle motion noise model;
the neural network training module substitutes the triaxial acceleration signal segments in the test set into the vehicle motion noise model one by one to obtain an output result corresponding to the electrocardio-motion noise for the vehicle motion noise model obtained through training, so that the output result corresponding to each group of triaxial acceleration signal segments in the test set is compared with the three-channel electrocardio-motion noise, and if the comparison result of a certain proportion of triaxial acceleration signal segments in the test set is less than or equal to a threshold value, the vehicle motion noise model is finally determined; otherwise, the electrocardiosignal simulation generator and the three-axis acceleration sensor are used for acquiring more three-channel static electrocardiosignal segments, three-channel moving electrocardiosignal segments and three-axis acceleration signal segments, and the three-channel moving electrocardiosignal segments and the three-axis acceleration signal segments are used for training as the input of the neural network training module by increasing the scale of the training set after passing through the signal acquisition module and the preprocessing module in the intelligent terminal.
2. The intelligent terminal of claim 1, wherein: the artificial neural network learning algorithm adopted by the neural network training module takes a proportional conjugate gradient method as an optimization direction.
3. The intelligent terminal of claim 1, wherein: the input layer of the neural network constructed by the neural network training module through initialization consists of 3 neurons, the hidden layer consists of 10 neurons, and the output layer consists of 3 neurons.
4. The intelligent terminal of claim 1, wherein: the expression of the neuron function g (z) in the neural network initially constructed by the neural network training module is as follows:
wherein: z is the argument of the function.
5. The intelligent terminal of claim 1, wherein: the intelligent terminal is an intelligent mobile phone, a tablet personal computer or a PC.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610113597.XA CN105796091B (en) | 2016-02-29 | 2016-02-29 | A kind of intelligent terminal for removing electrocardiosignal vehicle movement noise |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610113597.XA CN105796091B (en) | 2016-02-29 | 2016-02-29 | A kind of intelligent terminal for removing electrocardiosignal vehicle movement noise |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105796091A CN105796091A (en) | 2016-07-27 |
| CN105796091B true CN105796091B (en) | 2018-09-28 |
Family
ID=56465929
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610113597.XA Expired - Fee Related CN105796091B (en) | 2016-02-29 | 2016-02-29 | A kind of intelligent terminal for removing electrocardiosignal vehicle movement noise |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105796091B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108470158B (en) * | 2018-03-08 | 2020-05-12 | 华南理工大学 | Method for searching error minimum network computing structure for dynamic ECG data |
| LU100925B1 (en) * | 2018-09-10 | 2020-03-10 | Iee Sa | Removing noise caused by vehicular movement from sensor signals using Deep Neural Networks |
| CN113491513B (en) * | 2020-04-08 | 2023-06-30 | 华为技术有限公司 | Heart rhythm detection control method and terminal |
| CN112089412A (en) * | 2020-04-17 | 2020-12-18 | 中国人民解放军第二军医大学 | Method for performing algorithm denoising by using motion sensor in cooperation with electrocardiogram monitoring |
| CN115245334A (en) * | 2022-09-13 | 2022-10-28 | 北京昆迈医疗科技有限公司 | Magnetic field noise elimination method based on limb movement analysis |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104665849A (en) * | 2014-12-11 | 2015-06-03 | 西南交通大学 | Multi-physiological signal multi-model interaction-based high-speed railway dispatcher stress detecting method |
| CN105246400A (en) * | 2013-03-15 | 2016-01-13 | 佐尔医药公司 | ECG Noise Reduction System for Removing Vehicle Motion Artifacts |
| WO2016024081A1 (en) * | 2014-08-13 | 2016-02-18 | Saumarez Richard C | Improvement to analysing physiological electrograms |
-
2016
- 2016-02-29 CN CN201610113597.XA patent/CN105796091B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105246400A (en) * | 2013-03-15 | 2016-01-13 | 佐尔医药公司 | ECG Noise Reduction System for Removing Vehicle Motion Artifacts |
| WO2016024081A1 (en) * | 2014-08-13 | 2016-02-18 | Saumarez Richard C | Improvement to analysing physiological electrograms |
| CN104665849A (en) * | 2014-12-11 | 2015-06-03 | 西南交通大学 | Multi-physiological signal multi-model interaction-based high-speed railway dispatcher stress detecting method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105796091A (en) | 2016-07-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105769173B (en) | A kind of cardioelectric monitor system with electrocardiosignal noise removal function | |
| CN105796091B (en) | A kind of intelligent terminal for removing electrocardiosignal vehicle movement noise | |
| EP3692900B1 (en) | Artificial intelligence self-learning-based static electrocardiography analysis method and apparatus | |
| CN110420019B (en) | Deep regression heart rate estimation method for ballistocardiogram signals | |
| CN107378944B (en) | Multidimensional surface electromyographic signal artificial hand control method based on principal component analysis method | |
| US10806374B2 (en) | Noninvasive systems and methods for monitoring health characteristics | |
| Yoon et al. | Improvement of dynamic respiration monitoring through sensor fusion of accelerometer and gyro-sensor | |
| CN105534517A (en) | Method for removing vehicle motion noise in three-lead electrocardiosignal | |
| Chaudhuri et al. | Ambulation analysis in wearable ECG | |
| CN109907752A (en) | A kind of cardiac diagnosis and monitoring method and system of the interference of removal motion artifacts and ecg characteristics detection | |
| CN115299899B (en) | Multi-sensor-based activity identification and beat-to-beat blood pressure monitoring, analyzing and early warning system | |
| CN109446972B (en) | Gait recognition model establishing method, recognition method and device based on electromyographic signals | |
| CN111839506B (en) | Mental load detection method and device | |
| Zhang et al. | Upper limb motion capturing and classification for unsupervised stroke rehabilitation | |
| CN113116321A (en) | Non-invasive continuous blood pressure measuring system based on PSO-GRNN neural network | |
| Rehman et al. | Analytical survey of wearable sensors | |
| CN113520413B (en) | Lower limb multi-joint angle estimation method based on surface electromyogram signals | |
| Zou et al. | A multimodal fusion model for estimating human hand force: Comparing surface electromyography and ultrasound signals | |
| CN117813052A (en) | Method for generating synchronous electrocardiogram based on 2-lead asynchronous electrocardiogram | |
| Pombo et al. | Big data reduction using RBFNN: A predictive model for ECG waveform for eHealth platform integration | |
| Iadarola et al. | A new method for sEMG envelope detection from reduced measurements | |
| Gong et al. | Design and implementation of wearable dynamic electrocardiograph real-time monitoring terminal | |
| CN212186496U (en) | Simulation system for real-time compression sampling reconstruction of electromyographic signals | |
| Liu et al. | Semantic segmentation of qrs complex in single channel ecg with bidirectional lstm networks | |
| CN113261951B (en) | Sleeping posture identification method and device based on piezoelectric ceramic sensor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20190729 Address after: 312000, Zhejiang, Shaoxing City, Pingjiang Road, No. 328, ninth, one floor Patentee after: ZHEJIANG MEDZONE MEDICAL EQUIPMENT CO.,LTD. Address before: 312000, Shaoxing, Pingjiang Road, No. 328, ninth, two, Zhejiang Co-patentee before: ZHEJIANG MEDZONE MEDICAL EQUIPMENT CO.,LTD. Patentee before: ZHEJIANG MINGZHONG TECHNOLOGY Co.,Ltd. Co-patentee before: ZHEJIANG MEDZONE BIOMEDICAL MATERIAL AND DEVICE Research Institute |
|
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200918 Address after: Room 703, 7 / F, No. 3766, Nanhuan Road, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: ZHEJIANG MINGZHONG TECHNOLOGY Co.,Ltd. Address before: 312000, Zhejiang, Shaoxing City, Pingjiang Road, No. 328, ninth, one floor Patentee before: ZHEJIANG MEDZONE MEDICAL EQUIPMENT Co.,Ltd. |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180928 |