CN113712520A - Wearable arrhythmia detection device - Google Patents
Wearable arrhythmia detection device Download PDFInfo
- Publication number
- CN113712520A CN113712520A CN202111048228.4A CN202111048228A CN113712520A CN 113712520 A CN113712520 A CN 113712520A CN 202111048228 A CN202111048228 A CN 202111048228A CN 113712520 A CN113712520 A CN 113712520A
- Authority
- CN
- China
- Prior art keywords
- layer
- arrhythmia
- wearable
- acquisition unit
- detection device
- 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.)
- Pending
Links
- 206010003119 arrhythmia Diseases 0.000 title claims abstract description 71
- 230000006793 arrhythmia Effects 0.000 title claims abstract description 70
- 238000001514 detection method Methods 0.000 title claims abstract description 36
- 230000029058 respiratory gaseous exchange Effects 0.000 claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 23
- 238000013528 artificial neural network Methods 0.000 claims abstract description 21
- 230000008569 process Effects 0.000 claims abstract description 11
- 238000013527 convolutional neural network Methods 0.000 claims description 25
- 230000004913 activation Effects 0.000 claims description 20
- 238000002565 electrocardiography Methods 0.000 claims description 19
- 230000002028 premature Effects 0.000 claims description 19
- 230000000241 respiratory effect Effects 0.000 claims description 19
- 238000004891 communication Methods 0.000 claims description 18
- 238000012549 training Methods 0.000 claims description 13
- 230000001746 atrial effect Effects 0.000 claims description 12
- 230000002861 ventricular Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000011176 pooling Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 206010006578 Bundle-Branch Block Diseases 0.000 claims description 6
- 238000012952 Resampling Methods 0.000 claims description 5
- 230000000747 cardiac effect Effects 0.000 claims description 5
- 230000001667 episodic effect Effects 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 5
- 206010006580 Bundle branch block left Diseases 0.000 claims description 3
- 206010006582 Bundle branch block right Diseases 0.000 claims description 3
- 201000001715 left bundle branch hemiblock Diseases 0.000 claims description 3
- 210000002569 neuron Anatomy 0.000 claims description 3
- 201000007916 right bundle branch block Diseases 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims 1
- 230000008859 change Effects 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 24
- 230000005540 biological transmission Effects 0.000 description 8
- 238000012545 processing Methods 0.000 description 8
- 239000002184 metal Substances 0.000 description 5
- 208000019622 heart disease Diseases 0.000 description 4
- 230000002159 abnormal effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 108091006146 Channels Proteins 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005284 excitation Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000002847 impedance measurement Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000033764 rhythmic process Effects 0.000 description 2
- 210000001013 sinoatrial node Anatomy 0.000 description 2
- 208000014526 Conduction disease Diseases 0.000 description 1
- 206010019280 Heart failures Diseases 0.000 description 1
- 206010000891 acute myocardial infarction Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/0205—Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
- A61B5/024—Detecting, measuring or recording pulse rate or heart rate
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/08—Detecting, measuring or recording devices for evaluating the respiratory organs
-
- 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]
- A61B5/33—Heart-related electrical modalities, e.g. electrocardiography [ECG] specially adapted for cooperation with other devices
-
- 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]
- A61B5/332—Portable devices specially adapted therefor
-
- 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]
- A61B5/346—Analysis of electrocardiograms
- A61B5/349—Detecting specific parameters of the electrocardiograph cycle
- A61B5/352—Detecting R peaks, e.g. for synchronising diagnostic apparatus; Estimating R-R interval
-
- 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
-
- 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
- A61B5/725—Details of waveform analysis using specific filters therefor, e.g. Kalman or adaptive filters
-
- 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
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- 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
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
Abstract
The invention provides a wearable arrhythmia detection device, which comprises an electrocardiosignal acquisition unit, a respiration signal acquisition unit, a controller and a PC (personal computer), and is characterized in that the controller controls the electrocardiosignal acquisition unit and the respiration signal acquisition unit, and the PC processes the electrocardiosignal and the respiration signal so as to realize automatic detection of arrhythmia of a subject. The wearable arrhythmia detection device provided by the invention can be used for detecting the electrocardio change of a testee in real time, effectively utilizing the acquired electrocardio signals to extract the characteristics, and finally classifying arrhythmia by using the deep neural network, thereby realizing the purpose of wearable arrhythmia detection.
Description
Technical Field
The invention relates to the technical field of medical equipment, in particular to a wearable arrhythmia detection device.
Background
Along with the development of science and technology and the improvement of the living standard of people, the attention degree of people to the health condition is also being showing and is promoting, has more and more people hope to know the health condition of self to in time seek medical advice in finding the problem. The heart is a vital organ of a human body, however, because of the influence of epidemic situations, procedures for measuring electrocardio in a hospital are complicated, so that automatic detection related to heart diseases is also concerned widely, and an electrocardio sensor and a computer play a vital role in detection and diagnosis of heart diseases.
Arrhythmia refers to abnormal frequency and/or rhythm of heart beats caused by abnormal sinus node activation or activation generated outside the sinus node, such as by slow conduction, blocking or conduction through an abnormal channel, i.e., the origin of heart activity and/or conduction disorder. Genetic arrhythmia is mostly caused by gene channel mutation; acquired arrhythmias can be seen in various organic heart diseases, especially in the development of heart failure or acute myocardial infarction. Therefore, the method is particularly important for detecting arrhythmia, and can effectively detect and early warn various heart diseases.
Currently, most of known devices for detecting arrhythmia are large electrocardiosignal detection devices used in hospitals and the like, a subject is required to lie on a diagnosis and treatment bed in a test process, high requirements are imposed on test environment and the state of the subject, and judgment of arrhythmia depends on professional knowledge and experience of doctors, so that portable and automatic test cannot be realized.
The invention patent application with application publication number CN108937912A discloses an automatic heart rhythm market analysis method based on a deep neural network, which comprises the following steps: generating a multi-channel electrocardiogram sample by three sampling modes; splicing the obtained 600-dimensional electrocardiosignals along a second dimension, amplifying the electrocardiosignals from 600 × 1 dimensions to 600 × 3 dimensions, inputting the electrocardiosignals into a plurality of convolution layer units and LSTM layer units which are sequentially connected in series, wherein an entry layer is arranged between the convolution layer units and the LSTM layer units; the convolutional layer unit comprises a convolutional layer using one-dimensional convolution and an excitation unit operation and a pooling layer operation which are sequentially connected in series with the output end of the convolutional layer; the convolution layer is used for extracting the characteristics of the one-dimensional electrocardiosignals; the output of the LSTM layer unit is connected with a full connection layer of which the excitation unit is softmax in series; outputting; and learning parameters of the deep neural network, and automatically identifying the sample. The method has the defects that the data volume of a sample to be processed is large, the method for extracting the characteristics of the electrocardiosignal is complicated, the time consumption is long, and the arrhythmia detection efficiency of the system is influenced.
Disclosure of Invention
In order to solve the technical problems, the wearable arrhythmia detection device provided by the invention can detect the electrocardio change of a subject in real time by using the wearable arrhythmia monitoring device, effectively utilizes the acquired electrocardio signals to extract the characteristics, and finally classifies arrhythmia by using a deep neural network, thereby achieving the purpose of wearable arrhythmia detection.
The invention provides a wearable arrhythmia detection device, which comprises an electrocardiosignal acquisition unit, a respiration signal acquisition unit, a controller and a PC (personal computer), wherein the controller controls the electrocardiosignal acquisition unit and the respiration signal acquisition unit, and the PC processes the electrocardiosignal and the respiration signal so as to realize automatic detection of arrhythmia of a subject.
Preferably, the input end of the electrocardiosignal acquisition unit is connected to a three-lead electrocardio electrode, the input end of the respiratory signal acquisition unit is impedance between a positive electrode and a negative electrode of the electrocardiosignal, and the output ends of the electrocardiosignal acquisition unit and the respiratory signal acquisition unit are connected to the input end of the controller.
In any of the above aspects, preferably, the controller is configured to control acquisition of the cardiac and respiratory signals and communication of the transmission of the data.
In any of the above schemes, preferably, the apparatus further includes a wireless communication unit, the wireless communication unit is a bluetooth module, a transmitting end of the wireless communication unit is connected to the controller, a receiving end of the wireless communication unit is connected to the PC, and the PC displays the electrocardiographic and respiratory waveforms of the subject.
In any of the above aspects, preferably, the PC is used for display of the subject's cardiac and respiratory waveforms, data processing, and automatic classification of cardiac arrhythmias using a deep neural network.
In any of the above embodiments, preferably, the data processing is performed by filtering the cardiac signal and the respiratory signal using a butterworth filter.
In any of the above schemes, preferably, the automatic classification of the arrhythmia by using the deep neural network is to simultaneously use two models, namely a forward feedback neural network FFNN and a convolutional neural network CNN, to realize the automatic classification of the arrhythmia.
In any of the above schemes, preferably, the FFNN includes five layers, which are an input layer, a hidden layer 1, a hidden layer 2, a hidden layer 3 and an output layer, all layers are connected, the output layer uses a softmax activation function, and the other layers use a ReLU activation function.
In any of the above schemes, preferably, the FFNN adopts full connection between layers, and the forward calculation formula from the n-1 th layer to the n-th layer is
X(n)＝f(n)(W(n)X(n-1)+B(n))
The output layer adopts a softmax activation function, and the expression is
The other layer adopts a ReLU activation function, and the expression is
f(x)＝max(0,x)
Wherein W is a linear relationship between layers, f(n)As a function of layer n-1 to layer n, X(n-1)Is the output signal of the n-1 th layer, B is the bias, xiTo output an input signal for layer i, j refers to traversal through n neurons.
In any of the above schemes, preferably, the FFNN inputs 200ms ecg segments and outputs detected R-wave position information for detecting the R-wave position of the ecg signal, and further obtains the feature value RR of arrhythmia automatic classification by resampling calculation, where the formula is
RR＝Resample(ECGs[qrsm-2:qrsm],360)
Wherein ECGs represents the ECG buffer, qrsm-2The previous R-wave point position, qrs, representing the current R-wave pointmThe position of the next R-wave point after the current R-wave point is indicated.
In any of the above schemes, preferably, the CNN is configured to extract the electrocardiosignal features RR, implement training and testing on arrhythmia, implement feature transformation of the feature values RR in a manner that each convolutional layer and pooling layer are connected in series, and measure loss between the output calculated for the training sample and the output of the real training sample.
In any of the above schemes, preferably, the output of each convolutional layer unit of the CNN is connected in series with a pooling layer operation, and the loss function is a cross-entropy function.
In any of the above schemes, it is preferable that the output layer dimension of the CNN is 1 × 13, and the CNN mainly includes 13 categories of arrhythmia, which are normal heartbeat N, left bundle branch block L, right bundle branch block R, (nodal) junctional escape J, atrial escape E, atrial premature a, heterosis atrial premature a, junctional (junctional) premature J, supraventricular premature or ectopic beat S, ventricular premature V, ventricular escape E, ventricular fusion heartbeat F, and unclassified heartbeat Q.
In any of the above schemes, preferably, the three-lead electrocardiograph electrode is detachable from the wearing band, and the band is also detachable from the controller.
The invention provides a wearable arrhythmia detection device, which is used for detecting electrocardio and respiratory signals of a testee in real time, monitoring physiological indexes of the testee, displaying electrocardio and respiratory waveforms of the testee and realizing automatic classification of arrhythmia.
Drawings
Fig. 1 is a block diagram of a preferred embodiment of a wearable arrhythmia detection device in accordance with the present invention.
Fig. 2 is a schematic diagram of an embodiment of a deep neural network classification process in a wearable arrhythmia detection device according to the present invention.
Fig. 3 is a basic block diagram of a preferred embodiment of a wearable arrhythmia detection device according to the invention.
Detailed Description
The invention is further illustrated with reference to the figures and the specific examples.
Example one
As shown in fig. 1, a wearable arrhythmia detection device includes a three-lead electrocardiograph electrode 100, an electrocardiograph signal acquisition unit 110, a respiration signal acquisition unit 120, a controller 130, a PC 140, and a wireless communication unit 150.
The controller 130 controls the electrocardiosignal acquisition unit 110 and the respiration signal acquisition unit 120, and the PC 140 processes the electrocardiosignals and the respiration signals so as to realize automatic detection of arrhythmia of the subject.
The input end of the electrocardiosignal acquisition unit 120 is connected to the three-lead electrocardio-electrode 100, the input end of the respiration signal acquisition unit 130 is the impedance between the positive electrode and the negative electrode of the electrocardiosignal, and the output ends of the electrocardiosignal acquisition unit 120 and the respiration signal acquisition unit 130 are connected to the input end of the controller 140.
The controller 120 is used to control the acquisition of the cardiac and respiratory signals, as well as the transmission of data.
The wireless communication unit 150 is a bluetooth module, the transmitting end of the wireless communication unit 150 is connected to the controller 130, the receiving end of the wireless communication unit 150 is connected to the PC, and the ecg and respiration waveforms of the subject are displayed on the PC 140.
The PC 140 is used for displaying electrocardio and respiration waveforms of a subject, processing data and automatically classifying arrhythmia by using a deep neural network, wherein the data processing is to filter electrocardio signals and respiration signals by using a Butterworth filter, and the automatic classification of arrhythmia by using the deep neural network is to realize the automatic classification of arrhythmia by simultaneously using a forward feedback neural network FFNN model and a convolutional neural network CNN model.
The FFNN comprises five layers which are an input layer, a hidden layer 1, a hidden layer 2, a hidden layer 3 and an output layer, wherein the layers are fully connected, the output layer adopts a softmax activation function, and other layers adopt a ReLU activation function. The layers of the FFNN are all connected, and the forward calculation formula from the n-1 th layer to the n-th layer is
X(n)＝f(n)(W(n)X(n-1)+B(n))
The output layer adopts a softmax activation function, and the expression is
The other layer adopts a ReLU activation function, and the expression is
f(x)＝max(0,x)
Wherein W is a linear relationship between layers, f(n)As a function of layer n-1 to layer n, X(n-1)Is the output signal of the n-1 th layer, B is the bias, xiTo output an input signal for layer i, j refers to traversal through n neurons.
The input of the FFNN is 200ms electrocardio segments, the output is detected R wave position information which is used for realizing the detection of the R wave position of the electrocardiosignal, and then a characteristic value RR of automatic arrhythmia classification is obtained by utilizing resampling calculation, and the formula is
RR＝Resample(ECGs[qrsm-2:qrsm],360)
Wherein ECGs represents the ECG buffer, qrsm-2The previous R-wave point position, qrs, representing the current R-wave pointmThe position of the next R-wave point after the current R-wave point is indicated.
The CNN is used for training and testing arrhythmia after extracting electrocardiosignal features RR, realizing feature transformation of the feature value RR in a mode that each convolution layer is connected with a pooling layer in series, and measuring loss between output calculated by the training sample and real training sample output. The output of each convolutional layer unit of CNN is connected with a pooling layer operation in series, and the loss function selects the cross entropy function. The output layer dimension of CNN is 1 × 13, mainly comprising 13 classes of arrhythmia, namely normal heartbeat N, left bundle branch block L, right bundle branch block R, (episodic) junctional escape J, atrial escape E, atrial premature a, heterosis atrial premature a, junctional (junctional) premature J, supraventricular or ectopic beat S, ventricular premature V, ventricular escape E, ventricular fusion heartbeat F and unclassifiable heartbeat Q.
The three-lead electrocardio-electrode 100 can be detached from the wearing bandage, and the bandage can also be detached from the controller.
Example two
The wearable arrhythmia detection device provided by the invention specifically comprises an electrocardio and respiration signal acquisition unit, a controller and a PC:
the acquisition unit of the electrocardio and respiratory signals is used for contacting with a subject and acquiring the electrocardio and respiratory signals;
the controller is used for acquiring control signals and transmitting data;
the PC is used for receiving, processing and displaying the measured electrocardio and respiratory signals of the testee, and finally realizing automatic classification of arrhythmia.
The process of the controller for collecting the control signal and transmitting the data mainly comprises controlling an electrocardiosignal collecting chip and utilizing I2C, communication is carried out between the electrocardio acquisition chip and the controller, and communication between the controller and the PC is carried out by utilizing Bluetooth transmission.
The PC machine performs waveform display and data processing of the electrocardio and respiratory signals, and utilizes the deep neural network to realize automatic classification of arrhythmia.
The electrocardio acquisition unit comprises three electrocardio electrodes and a reference electrode, the input end of the electrocardio acquisition unit is contacted with the skin of the heart part of the human body, and the output end of the electrocardio acquisition unit is connected with the input end of the controller.
The respiration acquisition unit consists of an electrocardio-electrode A and a reference electrode, and the respiration rate is further obtained by calculating the impedance value between the electrocardio-electrode A and the reference electrode.
The respiratory impedance RbCalculating according to formula (1):
in the formula of UbIs an equivalent voltage applied to the human body; i isbIs equivalent current loaded to human body; u shape2Outputting a voltage for the impedance measurement circuit; rSThe resistance is a fixed value reference resistance in an impedance measurement circuit.
In the process of detecting the electrocardiosignal by the electrocardio acquisition unit, the electrocardio electrode adopts an electrode with small contact impedance with the skin of a human body and can be connected with an external circuit board by a lead; the conducting wire adopts a signal transmission conducting wire which has a shielding function, excellent conductivity and durability and does not influence the comfort degree of a wearer.
The wireless communication module adopts a Bluetooth module, the input end of the wireless communication module is connected to the controller, the receiving end of the Bluetooth module is connected to a PC (personal computer), and the PC is used for displaying the electrocardio and respiration waveforms and realizing the automatic detection of arrhythmia.
The electrocardio acquisition unit and the respiration acquisition unit form a whole and are connected with a bandage of a wearable garment in a metal button mode to form a detachable device.
The PC waveform display interface comprises a serial port setting interface, a data transmission interface, a waveform display interface, data storage and processing, and judgment and classification of arrhythmia by using a deep neural network.
The data preprocessing adopts three groups of Butterworth filters, and mainly comprises a low-pass filter with the cut-off frequency of 100Hz, a band-pass filter with the cut-off frequency of 49-51Hz and a high-pass filter with the cut-off frequency of 0.2 Hz.
As shown in fig. 2, the deep neural network mainly adopts two models, namely a forward feedback neural network FFNN model and a convolutional neural network CNN model:
the FFNN comprises five layers which are an input layer, a hidden layer 1, a hidden layer 2, a hidden layer 3 and an output layer, wherein the layers are all connected, the output layer adopts a softmax activation function, and other layers adopt a ReLU activation function. The input of the FFNN is 200ms of electrocardiographic segments, the output is detected R-wave position information, which is used to realize the detection of the R-wave position of the electrocardiographic signals, and then the characteristic value RR of the arrhythmia automatic classification is obtained by utilizing resampling calculation (the sequence of the ECG signals between the previous R-wave point and the next R-wave point of the current R-wave point being resampled to a fixed length (360), as shown in formula (2)).
RR＝Resample(ECGs[qrsm-2:qrsm],360) (2)
In which ECGs represents the ECG buffer, qrsm-2The previous R-wave point position, qrs, representing the current R-wave pointmThe position of the next R-wave point after the current R-wave point is indicated.
The CNN realizes the characteristic transformation of input signals by utilizing convolutional layers and pooling layers, the output end of each convolutional layer unit is connected with a pooling layer in series for operation, a cross entropy function is selected as a loss function, the dimension of the output layer of the CNN is 1 x 13, and the CNN mainly comprises 13 categories of arrhythmia, namely normal heartbeat N, left bundle branch conduction block L, right bundle branch conduction block R, (nodal) boundary escape J, atrial escape E, atrial premature A, heterosis atrial premature a, nodal (boundary) premature J, supraventricular premature or ectopic beat S, ventricular premature V, ventricular escape E, ventricular fusion heartbeat F and unsortendable heartbeat Q.
The CNN realizes a training set required in the arrhythmia classification process, and adopts an open source arrhythmia database provided in an MIT-BIH database.
Compared with the prior art, the invention has the beneficial effects that:
(1) the invention is a wearable device, which can detect the electrocardio and respiration signals of a testee in real time, monitor the physiological indexes of the testee, display the electrocardio and respiration waveforms of the testee and realize the automatic classification of arrhythmia;
(2) the wearable device is detachably designed, the electrocardio-electrode is connected with the wearable binding belt through the metal hidden button, and the wearable binding belt is connected with the controller through the metal hidden button;
(3) the communication transmission of the invention utilizes the Bluetooth module, reduces the use of wires, enlarges the activity range of the testee in the test process and reduces the influence on the actions of the testee.
EXAMPLE III
The arrhythmia detection device of the invention comprises an acquisition unit of electrocardiosignals and respiratory signals, a controller and a PC:
the acquisition unit of the electrocardiosignals and the respiratory signals is used for monitoring the electrocardiosignals and the respiratory signals, the input end of the acquisition unit is connected with the electrocardio-electrode, and the output end of the acquisition unit is connected with the input end of the controller;
the controller is used for realizing the control of the acquisition unit and the transmission of the acquired signals, and the data transmission between the controller and the PC is realized through the Bluetooth module;
the PC is used for acquiring waveform display, data processing and arrhythmia automatic classification of electrocardiosignals and respiratory signals.
The electrode arrangement mode and the design of the wearable binding belt are shown in figure 3, the electrocardio-electrode is connected with the wearable binding belt through a metal hidden button, the wearable binding belt is also connected with the controller through the metal hidden button, and the controller is kept to be positioned at the left side of the human body in the wearing process.
Because the electrocardiosignal is weak and is easily interfered by noise in the acquisition and transmission processes, the electrocardiosignal is required to be preprocessed after being acquired, and the noise is filtered on the premise of ensuring the integrity of the electrocardiosignal.
After preprocessing the electrocardiosignals, performing feature extraction and arrhythmia classification on the electrocardiosignals by using a deep neural network, wherein the deep neural network mainly comprises two neural network models: a feed-forward neural network FFNN and a convolutional neural network CNN. The FFNN is used for monitoring the position of the R wave of the electrocardiosignal, the input of the FFNN is 200ms of electrocardiosignal segments, and the output of the FFNN is the position information of the R wave. The layers of the FFNN are in full connection, and forward calculation formulas from the (n-1) th layer to the nth layer are shown as (3):
X(n)＝f(n)(W(n)X(n-1)+B(n)) (3)
wherein W is the linear relationship between layers and B is the bias.
The output layer adopts a softmax activation function, and the expression of the softmax activation function is shown as (4):
the other layers adopt a ReLU activation function, and the expression of the ReLU activation function is shown as (5):
f(x)＝max(0,x) (5)
after the R-wave position is detected by using FFNN, the ECG signal between the previous R-wave point and the next R-wave point of the current R-wave point of the characteristic RR of the electrocardiograph signal is calculated by resampling to a sequence with a fixed length (360), as shown in formula (6).
RR＝Resample(ECGs[qrsm-2:qrsm],360) (6)
In which ECGs represents the ECG buffer, qrsm-2The previous R-wave point position, qrs, representing the current R-wave pointmThe position of the next R-wave point after the current R-wave point is indicated.
The CNN is used for training and testing arrhythmia after extracting electrocardiosignal features RR, the feature transformation of RR is realized in a mode that each convolution layer is connected with a pooling layer in series, the loss between the output calculated for measuring a training sample and the output of a real training sample is measured, calculation and optimization are carried out in the CNN by using a loss function, namely, the loss value is expected to be minimized, and then program output is optimized.
For a better understanding of the present invention, the foregoing detailed description has been given in conjunction with specific embodiments thereof, but not with the intention of limiting the invention thereto. Any simple modifications of the above embodiments according to the technical essence of the present invention still fall within the scope of the technical solution of the present invention. In the present specification, each embodiment is described with emphasis on differences from other embodiments, and the same or similar parts between the respective embodiments may be referred to each other. For the system embodiment, since it basically corresponds to the method embodiment, the description is relatively simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The method, apparatus and system of the present invention may be implemented in a number of ways. For example, the methods and systems of the present invention may be implemented in software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order for the steps of the method is for illustrative purposes only, and the steps of the method of the present invention are not limited to the order specifically described above unless specifically indicated otherwise. Furthermore, in some embodiments, the present invention may also be embodied as a program recorded in a recording medium, the program including machine-readable instructions for implementing a method according to the present invention. Thus, the present invention also covers a recording medium storing a program for executing the method according to the present invention.
The description of the present invention has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to practitioners skilled in this art. The embodiment was chosen and described in order to best explain the principles of the invention and the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated.
Claims (10)
1. The utility model provides a wearing formula arrhythmia detection device, includes electrocardiosignal acquisition unit, breathing signal acquisition unit, controller and PC, its characterized in that, the controller is right electrocardiosignal acquisition unit with breathing signal acquisition unit controls, the PC handles electrocardiosignal and breathing signal to the realization is to the automated inspection of experimenter arrhythmia.
2. The wearable arrhythmia detection device of claim 1, wherein an input end of the electrocardiosignal acquisition unit is connected to a three-lead electrocardio-electrode, an input end of the respiration signal acquisition unit is impedance between a positive electrode and a negative electrode of the electrocardiosignal, and output ends of the electrocardiosignal acquisition unit and the respiration signal acquisition unit are connected to an input end of the controller.
3. The wearable arrhythmia detection device of claim 2, further comprising a wireless communication unit, wherein the wireless communication unit is a bluetooth module, a transmitting end of the wireless communication unit is connected to the controller, a receiving end of the wireless communication unit is connected to the PC, and the PC displays the ecg and respiration waveforms of the subject.
4. The wearable arrhythmia detection device of claim 3, wherein the PC is configured to display, data process, and utilize a deep neural network to automatically classify arrhythmia in the form of a display of the subject's cardiac and respiratory waveforms.
5. The wearable arrhythmia detection device of claim 2, wherein the automatic classification of arrhythmia using the deep neural network is an automatic classification of arrhythmia using both a Forward Feedback Neural Network (FFNN) and a Convolutional Neural Network (CNN) model.
6. The wearable arrhythmia detection device of claim 1, wherein the FFNN comprises five layers, namely an input layer, a hidden layer 1, a hidden layer 2, a hidden layer 3 and an output layer, wherein full connections are adopted among the layers, the output layer adopts a softmax activation function, and other layers adopt a ReLU activation function.
7. The wearable arrhythmia detection device of claim 1, wherein the FFNN employs full connectivity between layers, and the forward calculation formula from layer n-1 to layer n is
X(n)＝f(n)(W(n)X(n-1)+B(n))
The output layer adopts a softmax activation function, and the expression is
The other layer adopts a ReLU activation function, and the expression is
f(x)＝max(0,x)
Wherein W is a linear relationship between layers, f(n)As a function of layer n-1 to layer n, X(n-1)Is the output signal of the n-1 th layer, B is the bias, xiTo output an input signal for layer i, j refers to traversal through n neurons.
8. The wearable arrhythmia detection device of claim 7, wherein the FFNN has an input of 200ms ECG segments and an output of detected R-wave position information, and is used for detecting the R-wave position of the ECG signal, and further using resampling calculation to obtain the characteristic value RR for automatic arrhythmia classification, wherein the formula is
RR＝Resample(ECGs[qrsm-2:qrsm],360)
Wherein ECGs represents the ECG buffer, qrsm-2The previous R-wave point position, qrs, representing the current R-wave pointmThe position of the next R-wave point after the current R-wave point is indicated.
9. The wearable arrhythmia detection device of claim 8, wherein the CNN is configured to perform training and testing on the arrhythmia after extracting the ecg signal features RR, perform feature transformation on the feature values RR by connecting each convolutional layer and pooling layer in series, and measure loss between the output calculated for the training samples and the output of the real training samples.
10. The wearable arrhythmia detection device of claim 9, wherein the CNN has an output layer dimension of 1 x 13 and comprises 13 classifications of arrhythmia, namely, normal heartbeat N, left bundle branch block L, right bundle branch block R, (episodic) episodic escape J, atrial escape E, atrial premature a, heterotic atrial premature a, episodic (episodic) premature J, supraventricular premature or ectopic beat S, ventricular premature V, ventricular escape E, ventricular fused heartbeat F, and unclassified heartbeat Q.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111048228.4A CN113712520A (en) | 2021-09-08 | 2021-09-08 | Wearable arrhythmia detection device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111048228.4A CN113712520A (en) | 2021-09-08 | 2021-09-08 | Wearable arrhythmia detection device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113712520A true CN113712520A (en) | 2021-11-30 |
Family
ID=78682455
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111048228.4A Pending CN113712520A (en) | 2021-09-08 | 2021-09-08 | Wearable arrhythmia detection device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113712520A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102779234A (en) * | 2012-06-28 | 2012-11-14 | 中国科学院苏州纳米技术与纳米仿生研究所 | Electrocardiogram classified processing method and device |
| CN104799849A (en) * | 2015-05-19 | 2015-07-29 | 南京大学 | Portable sleep apnea syndrome monitoring method and device |
| CN108836316A (en) * | 2018-05-02 | 2018-11-20 | 湖北工业大学 | A kind of R wave of electrocardiosignal extracting method based on BP neural network |
| CN108932452A (en) * | 2017-05-22 | 2018-12-04 | 中国科学院半导体研究所 | Arrhythmia cordis beat classification method based on multiple dimensioned convolutional neural networks |
| CN213465043U (en) * | 2020-07-29 | 2021-06-18 | 西安慧脑智能科技有限公司 | Electrocardio respiration acquisition circuit, equipment and system |
-
2021
- 2021-09-08 CN CN202111048228.4A patent/CN113712520A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102779234A (en) * | 2012-06-28 | 2012-11-14 | 中国科学院苏州纳米技术与纳米仿生研究所 | Electrocardiogram classified processing method and device |
| CN104799849A (en) * | 2015-05-19 | 2015-07-29 | 南京大学 | Portable sleep apnea syndrome monitoring method and device |
| CN108932452A (en) * | 2017-05-22 | 2018-12-04 | 中国科学院半导体研究所 | Arrhythmia cordis beat classification method based on multiple dimensioned convolutional neural networks |
| CN108836316A (en) * | 2018-05-02 | 2018-11-20 | 湖北工业大学 | A kind of R wave of electrocardiosignal extracting method based on BP neural network |
| CN213465043U (en) * | 2020-07-29 | 2021-06-18 | 西安慧脑智能科技有限公司 | Electrocardio respiration acquisition circuit, equipment and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6625682B2 (en) | Oxygen saturation measurement method | |
| Merdjanovska et al. | Comprehensive survey of computational ECG analysis: Databases, methods and applications | |
| JP4280806B2 (en) | Patient monitoring system with non-invasive cardiac output monitoring | |
| CN108577830B (en) | User-oriented physical sign information dynamic monitoring method and system | |
| WO2019161608A1 (en) | Multi-parameter monitoring data analysis method and multi-parameter monitoring system | |
| US20090024044A1 (en) | Data recording for patient status analysis | |
| JP6843122B2 (en) | High / low frequency signal quality evaluation of ECG read signal | |
| CN103445767B (en) | The full-automatic autonomic nervous function detector of sensor monitoring interactive controlling | |
| CN110897631B (en) | Real-time pregnancy monitoring device and method | |
| JP2009028441A (en) | Biological information measuring deice, biological information display method and display program | |
| KR20070087787A (en) | Emotion classification system of real time bio signals and emotion classification method | |
| Lai et al. | A flexible multilayered dry electrode and assembly to single-lead ECG patch to monitor atrial fibrillation in a real-life scenario | |
| Orha et al. | Automated system for evaluating health status | |
| JP2023099105A (en) | Electrocardiogram measurement method and system using wearable device | |
| US9198594B2 (en) | TWA measuring apparatus and TWA measuring method | |
| Mukti et al. | Performance analysis of twelve lead ECG based on delivery distance using bluetooth communication | |
| Rodríguez-Jorge et al. | Internet of things-assisted architecture for QRS complex detection in real time | |
| CN1228290A (en) | Neonatal physiological parameter monitor | |
| CN113712520A (en) | Wearable arrhythmia detection device | |
| CN113288156B (en) | Method for generating electrocardiographic data of any lead visual angle | |
| CN114903445A (en) | Intelligent monitoring and early warning system for cardiovascular and cerebrovascular diseases | |
| Rocha et al. | Wearable computing for patients with coronary diseases: Gathering efforts by comparing methods | |
| CN109431493B (en) | Wearable body surface potential acquisition device based on distance segmentation weighting algorithm | |
| Huda | Handheld Electrocardiogram Design | |
| Xie | Biomedical Signal Processing: An ECG Application |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211130 |