CN114711725A - Sleep apnea detection method and device based on double attention mechanism - Google Patents
Sleep apnea detection method and device based on double attention mechanism Download PDFInfo
- Publication number
- CN114711725A CN114711725A CN202210393579.7A CN202210393579A CN114711725A CN 114711725 A CN114711725 A CN 114711725A CN 202210393579 A CN202210393579 A CN 202210393579A CN 114711725 A CN114711725 A CN 114711725A
- Authority
- CN
- China
- Prior art keywords
- feature
- module
- rpeak
- rri
- sleep apnea
- 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
- 201000002859 sleep apnea Diseases 0.000 title claims abstract description 58
- 238000001514 detection method Methods 0.000 title claims abstract description 31
- 230000007246 mechanism Effects 0.000 title claims abstract description 26
- 238000012545 processing Methods 0.000 claims abstract description 45
- 238000000034 method Methods 0.000 claims abstract description 38
- 230000015654 memory Effects 0.000 claims abstract description 26
- 230000002457 bidirectional effect Effects 0.000 claims abstract description 17
- 238000000605 extraction Methods 0.000 claims abstract description 15
- 238000007781 pre-processing Methods 0.000 claims abstract description 12
- 238000013528 artificial neural network Methods 0.000 claims abstract description 5
- 238000013527 convolutional neural network Methods 0.000 claims description 26
- 230000009977 dual effect Effects 0.000 claims description 18
- 208000008784 apnea Diseases 0.000 claims description 9
- 238000011176 pooling Methods 0.000 claims description 8
- 238000011156 evaluation Methods 0.000 claims description 7
- 230000005284 excitation Effects 0.000 claims description 7
- 206010021079 Hypopnoea Diseases 0.000 claims description 6
- 238000001125 extrusion Methods 0.000 claims description 6
- 230000000747 cardiac effect Effects 0.000 claims description 4
- 230000007787 long-term memory Effects 0.000 claims description 4
- 230000006403 short-term memory Effects 0.000 claims description 4
- 230000006870 function Effects 0.000 description 14
- 238000010586 diagram Methods 0.000 description 9
- 238000004590 computer program Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000000241 respiratory effect Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 239000003550 marker Substances 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000029058 respiratory gaseous exchange Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 208000007590 Disorders of Excessive Somnolence Diseases 0.000 description 1
- 206010019280 Heart failures Diseases 0.000 description 1
- 206010020772 Hypertension Diseases 0.000 description 1
- 206010041349 Somnolence Diseases 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 208000029078 coronary artery disease Diseases 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000035475 disorder Diseases 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000000422 nocturnal effect Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000036387 respiratory rate Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 238000012795 verification 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
- A61B5/48—Other medical applications
- A61B5/4806—Sleep evaluation
- A61B5/4818—Sleep apnoea
-
- 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
- A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
- A61B5/346—Analysis of electrocardiograms
-
- 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
- A61B5/7264—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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Cardiology (AREA)
- Veterinary Medicine (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Theoretical Computer Science (AREA)
- Medical Informatics (AREA)
- Heart & Thoracic Surgery (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Psychiatry (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Fuzzy Systems (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a sleep apnea detection method and device based on a double attention mechanism, wherein the method comprises the following steps: acquiring an original electrocardiosignal, and performing data preprocessing on the original electrocardiosignal to obtain an RRI and an Rpeak; performing feature extraction on the RRI and the Rpeak through a one-dimensional convolution neural network module to obtain a first feature and a second feature; then, carrying out weight operation processing on the first characteristic and the second characteristic through a first attention module to obtain a third characteristic; then, the context dependency relationship of the third characteristic is captured through the bidirectional long-short term memory network module to obtain a fourth characteristic; then, performing weight operation processing on the first feature and the fourth feature through a second attention module to obtain a fifth feature; and finally, carrying out classification processing on the fifth features through a classifier to obtain a classification result. The invention can comprehensively consider the influence of electrocardiosignals on sleep apnea detection, realizes high-efficiency and accurate sleep apnea detection, and can be widely applied to the technical field of medical data detection.
Description
Technical Field
The invention relates to the technical field of medical data detection, in particular to a sleep apnea detection method and device based on a double attention mechanism.
Background
Sleep Apnea (SA) is a sleep-related breathing disorder characterized by repeated complete or incomplete cessation of respiratory airflow during sleep. It is estimated that about 10 million people worldwide suffer from this disease. In the adult population, over 14% of men and 5% of women suffer from sleep apnea. However, the vast majority of sleep apnea patients are not diagnosed and treated in a timely manner. Failure to diagnose and treat sleep apnea in a timely manner can lead to daytime sleepiness, hypertension, coronary heart disease, heart failure, and other complications. Timely diagnosis and treatment of sleep apnea is important to prevent its complications.
Polysomnography (PSG) is an internationally recognized and effective sleep apnea detection instrument. Polysomnography is a nocturnal recording of a patient and uses sensors attached to the body to measure various physiological signals, such as electroencephalogram (EEG), Electromyogram (EMG), Electrocardiogram (ECG), and Electrooculogram (EOG), for monitoring respiratory and other physiological changes in a breathing patient. When the polysomnography is used for detection, a patient needs to wear a large number of wires and electrodes in a hospital for sleep signal acquisition, which may bring uncomfortable experience and high cost. Furthermore, after collecting polysomnography data, physicians need to spend a great deal of time viewing and analyzing the data to make a diagnosis. Therefore, although polysomnography is the gold standard for diagnosing sleep apnea, a simple and inexpensive alternative is needed.
Electrocardiograms are easy to acquire and are one of the most physiologically relevant signals in SA events. Many studies have shown that detecting SA using single lead ECG signals is a low cost practical alternative. In recent years, many researchers have applied deep learning methods to detect data apnea from electrocardiographic signals.
In order to automatically detect the SA by using the electrocardiographic signal, some researchers mainly adopt a traditional machine learning method based on feature engineering, and need to design artificial features and select a proper classifier. For example, some methods extract 24 features from the R-peak interval of the ECG signal and 8 features from the ECG-derived respiratory signal, and segment the temporal correlation of the signals to create a model for SA detection. Some methods use a Support Vector Machine (SVM) to detect sleep apnea using heart rate and respiratory rate signals extracted from the cardiac electrical signal. However, these methods often rely heavily on the prior knowledge of experts in the feature extraction process, and the detection performance is greatly limited.
Disclosure of Invention
In view of this, embodiments of the present invention provide a sleep apnea detecting method and apparatus based on a dual attention mechanism, which can comprehensively consider the influence of an electrocardiographic signal on sleep apnea detection, and implement efficient and accurate sleep apnea detection.
In a first aspect, an embodiment of the present invention provides a sleep apnea detecting method based on a dual attention mechanism, including:
acquiring an original electrocardiosignal, and performing data preprocessing on the original electrocardiosignal to obtain an RRI and an Rpeak;
performing feature extraction on the RRI and the Rpeak through a one-dimensional convolution neural network module to obtain a first feature and a second feature;
performing weight operation processing on the first feature and the second feature through a first attention module to obtain a third feature;
acquiring a fourth characteristic by performing context dependency relationship capture processing on the third characteristic through a bidirectional long-short term memory network module;
performing weight operation processing on the first feature and the fourth feature through a second attention module to obtain a fifth feature;
and classifying the fifth features through a classifier to obtain a classification result.
Optionally, the acquiring the original electrocardiographic signal, and performing data preprocessing on the original electrocardiographic signal to obtain an RRI and an Rpeak includes:
acquiring an original electrocardiosignal;
extracting a mark segment and a segment adjacent to the mark segment according to the original electrocardiosignal to obtain 5-min-RRI, 5-min-Rpeak and 1-min-RRI-Rpeak;
carrying out length unified processing on the 5-min-RRI, the 5-min-Rpeak and the 1-min-RRI-Rpeak by an interpolation technology;
wherein the 5-min-RRI represents a RRI of 5 minutes in length, the 5-min-Rpeak represents a Rpeak of 5 minutes in length, and the 1-min-RRI-Rpeak represents a RRI and a Rpeak of 1 minute in length.
Optionally, the performing, by the one-dimensional convolutional neural network module, feature extraction on the RRI and the Rpeak to obtain a first feature and a second feature includes:
performing feature extraction on the 5-min-RRI and the 5-min-Rpeak through a first one-dimensional convolutional neural network to obtain F'rAnd F'p(ii) a Pairing said F 'in channel dimension'rAnd F'pSplicing to obtain a first characteristic;
performing feature extraction on the 1-min-RRI-Rpeak through a second one-dimensional convolutional neural network to obtain a second feature;
the first one-dimensional convolutional neural network and the second one-dimensional convolutional neural network respectively comprise three one-dimensional convolutional layers and two one-dimensional maximum pooling layers.
Optionally, the obtaining a third feature by fusing the first feature and the second feature through the first attention module includes:
performing extrusion processing and excitation processing on the first characteristic to obtain a first weight;
and carrying out scale operation on the first weight and the second characteristic to obtain a third characteristic.
Optionally, the obtaining a fourth feature by performing context dependency capture processing on the third feature through the bidirectional long-term and short-term memory network module includes:
adjusting, adding and deleting information flow through an LSTM unit, and finishing context dependency relationship capturing processing according to the adjusting, adding and deleting processing;
the bidirectional long-short term memory network module comprises a forward LSTM unit and a backward LSTM unit, and the LSTM unit comprises an input gate, an output gate and a forgetting gate.
Optionally, the performing, by the second attention module, a weight operation on the first feature and the fourth feature to obtain a fifth feature includes:
performing extrusion processing and excitation processing on the first features to obtain a second weight;
and carrying out scale operation on the second weight and the fourth feature to obtain a fifth feature.
Optionally, the method further comprises:
an evaluation index of sleep apnea detection and an apnea hypopnea index are determined.
In a second aspect, an embodiment of the present invention provides a sleep apnea detecting apparatus based on a dual attention mechanism, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring an original electrocardiosignal and carrying out data preprocessing on the original electrocardiosignal to obtain RRI and Rpeak;
the second module is used for extracting the features of the RRI and the Rpeak through the one-dimensional convolutional neural network module to obtain a first feature and a second feature;
a third module, configured to perform weight operation on the first feature and the second feature through the first attention module to obtain a third feature;
a fourth module, configured to perform context dependency capture processing on the third feature through a bidirectional long-term and short-term memory network module to obtain a fourth feature;
a fifth module, configured to perform weight operation on the first feature and the fourth feature through the second attention module to obtain a fifth feature;
and the sixth module is used for carrying out classification processing on the fifth characteristic through a classifier to obtain a classification result.
In a third aspect, an embodiment of the present invention provides an electronic device, including a processor and a memory;
the memory is used for storing programs;
the processor executes the program to implement the method according to the first aspect of the embodiment of the present invention.
In a fourth aspect, embodiments of the present invention provide a computer-readable storage medium storing a program for execution by a processor to implement a method as described above.
The embodiment of the invention also discloses a computer program product or a computer program, which comprises computer instructions, and the computer instructions are stored in a computer readable storage medium. The computer instructions may be read by a processor of a computer device from a computer-readable storage medium, and the computer instructions executed by the processor cause the computer device to perform the foregoing method.
The method comprises the steps of firstly obtaining an original electrocardiosignal, and carrying out data preprocessing on the original electrocardiosignal to obtain an RRI and an Rpeak; then, performing feature extraction on the RRI and the Rpeak through a one-dimensional convolution neural network module to obtain a first feature and a second feature; then, carrying out weight operation processing on the first feature and the second feature through a first attention module to obtain a third feature; then, a context dependency relationship is captured and processed on the third characteristic through a bidirectional long-short term memory network module to obtain a fourth characteristic; then, carrying out weight operation processing on the first characteristic and the fourth characteristic through a second attention module to obtain a fifth characteristic; and finally, carrying out classification processing on the fifth features through a classifier to obtain a classification result. According to the invention, the channel characteristics are accurately focused by the double-focusing module, so that important information is extracted from the signal segment of the electrocardio marker, the important channel in the fused characteristics is determined by utilizing the information, the context dependence relationship in the electrocardio characteristics is obtained by the bidirectional long-short term memory network module, and the influence of adjacent electrocardio signals on sleep apnea detection can be fully considered.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic flowchart of a sleep apnea detecting method based on a dual attention mechanism according to an embodiment of the present invention;
FIG. 2 is a schematic representation of RRI and Rpeak of an electrocardiogram;
FIG. 3 is a flowchart illustrating sleep apnea detection according to an embodiment of the present invention;
fig. 4 is a schematic diagram of a flow of extracting RRI and Rpeak data according to an embodiment of the present invention;
FIG. 5 is a schematic architectural diagram of a DANet according to an embodiment of the present invention;
fig. 6 is a detailed structural parameter diagram of a one-dimensional convolutional neural network module according to an embodiment of the present invention;
FIG. 7 is a schematic structural diagram of LSTM and BilSTM provided in an embodiment of the present invention;
FIG. 8 is a block diagram of a dual attention module provided in accordance with an embodiment of the present invention;
fig. 9 is a schematic structural diagram of a classifier according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In a first aspect, referring to fig. 1, an embodiment of the present invention provides a sleep apnea detection method based on a dual attention mechanism, including:
acquiring an original electrocardiosignal, and performing data preprocessing on the original electrocardiosignal to obtain an RRI and an Rpeak;
performing feature extraction on the RRI and the Rpeak through a one-dimensional convolution neural network module to obtain a first feature and a second feature;
performing weight operation processing on the first feature and the second feature through a first attention module to obtain a third feature;
acquiring a fourth characteristic by performing context dependency relationship capture processing on the third characteristic through a bidirectional long-short term memory network module;
performing weight operation processing on the first feature and the fourth feature through a second attention module to obtain a fifth feature;
and classifying the fifth features through a classifier to obtain a classification result.
Optionally, the acquiring the original electrocardiographic signal, and performing data preprocessing on the original electrocardiographic signal to obtain an RRI and an Rpeak includes:
acquiring an original electrocardiosignal;
extracting a mark segment and a segment adjacent to the mark segment according to the original electrocardiosignal to obtain 5-min-RRI, 5-min-Rpeak and 1-min-RRI-Rpeak;
carrying out length unified processing on the 5-min-RRI, the 5-min-Rpeak and the 1-min-RRI-Rpeak by an interpolation technology;
wherein the 5-min-RRI represents a RRI of 5 minutes in length, the 5-min-Rpeak represents a Rpeak of 5 minutes in length, and the 1-min-RRI-Rpeak represents a RRI and a Rpeak of 1 minute in length.
Optionally, the performing, by the one-dimensional convolutional neural network module, feature extraction on the RRI and the Rpeak to obtain a first feature and a second feature includes:
performing feature extraction on the 5-min-RRI and the 5-min-Rpeak through a first one-dimensional convolutional neural network to obtain F'rAnd F'p(ii) a Pairing said F 'in channel dimension'rAnd F'pSplicing to obtain a first characteristic;
performing feature extraction on the 1-min-RRI-Rpeak through a second one-dimensional convolutional neural network to obtain a second feature;
the first one-dimensional convolutional neural network and the second one-dimensional convolutional neural network respectively comprise three one-dimensional convolutional layers and two one-dimensional maximum pooling layers.
Optionally, the obtaining a third feature by fusing the first feature and the second feature through the first attention module includes:
performing extrusion processing and excitation processing on the first characteristic to obtain a first weight;
and carrying out scale operation on the first weight and the second characteristic to obtain a third characteristic.
Optionally, the obtaining a fourth feature by performing context dependency capture processing on the third feature through the bidirectional long-term and short-term memory network module includes:
adjusting, adding and deleting information flow through an LSTM unit, and finishing context dependency relationship capturing processing according to the adjusting, adding and deleting processing;
the bidirectional long-short term memory network module comprises a forward LSTM unit and a backward LSTM unit, and the LSTM unit comprises an input gate, an output gate and a forgetting gate.
Optionally, the performing, by the second attention module, a weight operation on the first feature and the fourth feature to obtain a fifth feature includes:
performing extrusion processing and excitation processing on the first features to obtain a second weight;
and performing scale operation on the second weight and the fourth feature to obtain a fifth feature.
Optionally, the method further comprises:
an evaluation index of sleep apnea detection and an apnea hypopnea index are determined.
In a second aspect, an embodiment of the present invention provides a sleep apnea detecting apparatus based on a dual attention mechanism, including:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring an original electrocardiosignal and carrying out data preprocessing on the original electrocardiosignal to obtain RRI and Rpeak;
the second module is used for extracting the features of the RRI and the Rpeak through the one-dimensional convolutional neural network module to obtain a first feature and a second feature;
a third module, configured to perform weight operation on the first feature and the second feature through the first attention module to obtain a third feature;
a fourth module, configured to perform context dependency capture processing on the third feature through the bidirectional long-short term memory network module to obtain a fourth feature;
a fifth module, configured to perform weight operation on the first feature and the fourth feature through the second attention module to obtain a fifth feature;
and the sixth module is used for carrying out classification processing on the fifth characteristics through the classifier to obtain a classification result.
The content of the method embodiment of the present invention is applicable to the apparatus embodiment, the functions specifically implemented by the apparatus embodiment are the same as those of the method embodiment, and the beneficial effects achieved by the apparatus embodiment are also the same as those achieved by the method.
Another aspect of the embodiments of the present invention further provides an electronic device, including a processor and a memory;
the memory is used for storing programs;
the processor executes the program to implement the method as described above.
The contents of the embodiment of the method of the present invention are all applicable to the embodiment of the electronic device, the functions specifically implemented by the embodiment of the electronic device are the same as those of the embodiment of the method, and the beneficial effects achieved by the embodiment of the electronic device are also the same as those achieved by the method.
Yet another aspect of the embodiments of the present invention provides a computer-readable storage medium, which stores a program, which is executed by a processor to implement the method as described above.
The contents of the embodiment of the method of the present invention are all applicable to the embodiment of the computer-readable storage medium, the functions specifically implemented by the embodiment of the computer-readable storage medium are the same as those of the embodiment of the method described above, and the advantageous effects achieved by the embodiment of the computer-readable storage medium are also the same as those achieved by the method described above.
The embodiment of the invention also discloses a computer program product or a computer program, which comprises computer instructions, and the computer instructions are stored in a computer readable storage medium. The computer instructions may be read by a processor of a computer device from a computer-readable storage medium, and the computer instructions executed by the processor cause the computer device to perform the foregoing method.
The present invention is further illustrated in detail below with reference to specific examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Firstly, the invention provides a deep learning model which fully utilizes electrocardiosignals. (1) The inputs to the model include the R peak Spacings (RRIs) and R peak volts (Rpeak) of adjacent segments and the RRIs and Rpeak of the labeled segments. The information provided by adjacent electrocardiosignals is utilized and two characteristics of RRI and Rpeak are fused. The interval of the R peak and the amplitude of the R peak of the electrocardiogram are shown in FIG. 2. (2) A dual attention module is proposed that not only extracts important information from the signal segments of the cardiac markers, but also uses this information to determine which channels of the fused features are more important and to assign corresponding weights. (3) The superiority of the method is verified through a generalization experiment.
In the present invention, the dataset PhysioNet Apnea-ECG dataset, which is publicly available from philips university, is used. The sampling frequency of the electrocardiogram signal is 100 hz. The data set segment-marks the electrocardiographic recordings, each for 1 minute, and marks these data segments as normal or sleep apnea. The total number of electrocardiographic recordings was 70, including two sets, a training data set containing 35 recordings (30% split for validation) and a test data set containing 35 recordings. In addition we collected 140 real data from one hospital.
The working flow of the dual attention network (DANet) SA detection proposed by the present invention is shown in fig. 3. First, the Hamilton algorithm is used to extract the R peak of the original electrocardiographic signal, and the R peak intervals (RRIs) and the R peak voltage value (Rpeak) are calculated. Then, the RRIs and Rpeak of the marked segment electrocardiosignals and the RRIs and Rpeak of each 2 segments electrocardiosignals adjacent to the marker segment electrocardiosignals are used as the input of the DANet. Finally, we train the DANet and then automatically detect and output whether each segment is normal or sleep apnea. At the same time, it is also determined whether each record is from a sleep apnea patient.
Specifically, the following describes the technical scheme of the invention in detail:
(1) data pre-processing
The r peak of the electrocardiogram signal was detected by the Hamilton algorithm, recorded as Rpeak, and the RRI was calculated. As the length of the electrocardiosignal segment is different and the information of the adjacent segment can influence the detection accuracy of the sleep apnea, as shown in figure 4, in the invention, 5-min-RRI, 5-min-Rpeak and 1-min-RRI-Rpeak are respectively extracted, so that the important information of the marked segment and the adjacent segment is combined to improve the detection performance. After filtering the extracted RRI by using a median filter, unifying the lengths of all the RRIs and Rpeak by using a cubic interpolation technology, and conveniently inputting the RRIs and the Rpeak into a model.
(2) Deep learning model DANet
The input to the model consists of 3 parts: 5-min-RRI, 5-min-Rpeak and 1-min-RRI-Rpeak, and is recorded as input ═ Fr，Fp，Frp]。
Fig. 5 shows the architecture of a DANet of the present invention, which consists of four parts: the system comprises a one-dimensional convolutional neural network module, a bidirectional long-time and short-time memory network module, a double attention module and a classifier module.
And a one-dimensional convolutional neural network (1D-CNN) module. 1D-CNN is intended to extract useful features from the input. The DANet contains two 1D-CNN, 1D-CNN-1 and 1D-CNN-2. The detailed structures and parameters of 1D-CNN-1 and 1D-CNN-2 are shown in FIG. 6, in which FIG. 6(a) is 1D-CNN-1 structure and parameters, and FIG. 6(b) is 1D-CNN-2 structure and parameters. Extracting 1D-CNN-1 to obtain 5-min-RRI FrIs characterized by the sum of 5-min-Rpeak FpAnd the two features are spliced on the channel dimension to obtain a fusion feature Fu. 1-min-RRI-Rpeak F is extracted from 1D-CNN-2rpThe method is characterized in that. Both 1D-CNN-1 and 1D-CNN-2 contained 3 one-dimensional convolutional layers (Conv1D) and 2 one-dimensional max pooling layers (Maxpooling1D), each followed by a ReLu function. The effect of the convolutional layer is to extract features. And the application of the nonlinear activation function ReLU to the convolution result is beneficial to forming the sparsity of the network and reducing the interdependence among the parameters. The role of the pooling layer is to reduce computational overhead.
1D-CNN-1. The input to 1D-CNN-1 is 5-min-RRI FrAnd 5-min-Rpeak FpBoth are 900 x 1 in size and both use an additional two adjacent segments as input. Input FrAnd FpThrough 1D-CNN-1 (parameter is denoted as theta)1) Obtaining output sum, and splicing to obtain FuAs in formulas (1), (2) and (3):
F′r＝f(Fr，θ1)#(1)
F′p＝f(Fp，θ1)#(2)
Fu＝[F′r，F′p]#(3)
1D-CNN-2 is different from 1D-CNN-1 in 1-min-RRI-Rpeak FrpIs the only input to 1D-CNN-2, 180X 2 in length. FrpF 'can be obtained through 1D-CNN-2'rpAs in equation (4):
F′rp＝f(Frp，θ2)#(4)
a Bidirectional Long-Short Term Memory Network (BilSTM) module. LSTM (Long-short-term Memory Network, LSTM for short) can solve the problems of gradient disappearance and Long-term dependence. Because the electrocardiographic data is not independently segmented but continuous. The electrocardiosignal of the current moment to be tested can be related to the signals of the previous moment and the next moment. Accordingly, BilSTM may be used to obtain contextual dependencies in electrocardiographic features. The structures of LSTM and BiLSTM are shown in fig. 7, where fig. 7(a) is a LSTM configuration diagram and fig. 7(b) is a BiLSTM configuration diagram. The BilSTM consists of a forward LSTM network and a reverse LSTM network, both of which comprise the same number of LSTM units, and the number of the LSTM units is set to be 16. The LSTM unit consists of an input gate, an output gate, and a forgetting gate that can add or delete information by adjusting the flow of information in the unit, thereby enabling the unit to capture information in a time stream.
A dual attention module. The detailed structure of the dual attention module is shown in fig. 8. The double attention mechanism is a mechanism widely applied in the field of deep learning at present. The attention mechanism is essentially to allocate resources according to the importance of the object, and the weight is the resource that needs to be allocated. Among them, the channel attention mechanism is a simple and effective attention mechanism. The channel attention mechanism performs a series of operations on channel dimensions, so that the model can focus on channel features containing important information. The invention adopts a one-dimensional channel attention mechanism.
The conventional channel attention mechanism learns features from itself only and gives channel weights to itself. In the invention, a one-dimensional channel is designed to be noticed, and a 1-minute long signal and a 5-minute long signal are combined for learning. Learning from the 1 minute long signal features, obtaining channel weights, and then assigning them to the 5 minute long signal features, a dual attention mechanism module consisting of channel-attention-1 and channel-attention-2 is proposed. Channel-annotation-1 was applied to the 1D-CNN module extracted features and channel-annotation-2 was applied to the BilSTM module extracted features.
channel-association-1. The input of channel-attribute-1 is feature F of 1D-CNN-1 fusionuAnd 1D-CNN-2 extracted feature F'rp. First, by squeezing (Squeeze) the feature fsqEncoding features on the channel into global features is achieved using a one-dimensional global average pooling operation. Then, the pass parameter is θex1Excitation operation fex1Capturing the relationship between channels, given by FuWeight of lambda1This is achieved by two fully connected layers. Finally, the scaling operation fscale1Will weight λ1And feature FuMultiplying to obtain new characteristic Fat1As in equations (5) and (6):
λ1＝fex1(fsq(F′rp；)；θex1)#(5)
wherein i represents a feature FuI ranges from 1 to 2c,
channel-attention-2. The input to channel-anchorage-2 is F'rpAnd features F extracted by the BilSTM moduleb. Similarly, F 'is first paired'rpPerforming a squeezing and energizing operation to obtain a distribution FbChannel weight λ of2Then the weight λ2And feature FbMultiplying to obtain a characteristic Fat2As in equations (7), (8) and (9):
Fb＝f(Fat1，θb)#(7)
λ2＝fex2(fsq(F′rp；)；θex2)#(8)
wherein, thetabRepresenting parameters of the BilSTM module, i ranging from 1 to c, c representing the characteristic FbThe number of the channels of (a) is,
and (4) a classifier. As shown in fig. 9, the classifier used in the present invention comprises a one-dimensional global average pool layer and a full-connection layer, wherein the activation function used in the full-connection layer is a Sigmoid function. The role of the pooling layer is to sparsely obtain features, reducing computational overhead. Last layer of deep neural networkThe features after convolution, excitation, pooling and other operations are connected in series by the full connection layer to serve as a voting value of classification, and a classification result is finally obtained. The input to the classifier is the feature F of the output of the dual attention moduleat2. The output of the classifier is Normal or Sleep Apnea (SA).
(3) Evaluation index
There are two phases of SA detection. The first stage, detecting whether the EEG signal segment with one minute of duration is SA. And in the second stage, whether the whole electroencephalogram signal record is SA or not is detected, and whether the electroencephalogram signal is from a sleep apnea patient or not can be judged. In the invention, the two SA detection evaluation indexes mainly comprise: accuracy (Acc), sensitivity (sensity, Sens), specificity (Spec), F1Value and model assessment index auc (area under the curve). Each evaluation index was calculated as follows:
wherein TP, TN, FP and FP represent "true positive", "true negative", "false positive" and "false negative", respectively. The invention takes the sleep apnea as Positive class and the normal as Negative class.
The performance metrics for each record included accuracy, sensitivity, specificity, Pearson correlation coefficient (Corr). Corr is the correlation coefficient between the true and predicted values of the AHI. Where AHI refers to the Apnea Hypopnea Index, which is an abbreviation for the English Apnea-Hypopnea Index. AHI refers to the number of apneas plus hypopneas per hour of sleep. It is actually the number of apneas divided by the number of nighttime hours of sleep. This is a widely used sleep apnea evaluation index. If the AHI is greater than or equal to 5, the subject suffers from sleep apnea. AHI is calculated as follows:
wherein, T represents the total number of 1 minute segment in one electroencephalogram signal record, and N represents the number of 1 minute segment with the label of SA in the electroencephalogram signal record.
(4) Experimental verification
In order to prove effectiveness, the algorithm provided by the invention is compared with the prior machine learning methods (HMM-SVM, LS-SVM and the like) and the deep learning-based methods (LeNet-5, LSTM, SE-MSCNN and the like) in a public data set PhysioNetApnea-ECG, wherein the performance comparison of the sleep apnea recognition of one-minute electrocardiosignal segments is shown in the table 1, and the performance comparison of the sleep apnea recognition of each electrocardiosignal record is shown in the table 2.
TABLE 1
TABLE 2
In the invention, a new deep learning technology-based algorithm DANet is developed to detect SA. The proposed DANet uses 1D-CNN and BiLSTM to learn the features for the 1-minute and 5-minute segments, respectively. And the 1 minute long signature and the 5 minute long signature were fused by a double attention mechanism. The detection accuracy for each segment and each record was 91.34% and 100%, respectively.
In alternative embodiments, the functions/acts noted in the block diagrams may occur out of the order noted in the operational illustrations. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality/acts involved. Furthermore, the embodiments presented and described in the flow charts of the present invention are provided by way of example in order to provide a more thorough understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and in which sub-operations described as part of larger operations are performed independently.
Furthermore, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise stated to the contrary, one or more of the described functions and/or features may be integrated in a single physical device and/or software module, or one or more functions and/or features may be implemented in a separate physical device or software module. It will also be appreciated that a detailed discussion of the actual implementation of each module is not necessary for an understanding of the present invention. Rather, the actual implementation of the various functional modules in the apparatus disclosed herein will be understood within the ordinary skill of an engineer, given the nature, function, and internal relationship of the modules. Accordingly, those skilled in the art can, using ordinary skill, practice the invention as set forth in the claims without undue experimentation. It is also to be understood that the specific concepts disclosed are merely illustrative of and not intended to limit the scope of the invention, which is defined by the appended claims and their full scope of equivalents.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The logic and/or steps represented in the flowcharts or otherwise described herein, such as an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution apparatus, device, or device (e.g., a computer-based apparatus, processor-containing apparatus, or other device that can fetch the instructions from the instruction execution apparatus, device, or device and execute the instructions). For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution apparatus, device, or apparatus.
More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by suitable instruction execution devices. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the invention have been shown and described, it will be understood by those of ordinary skill in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. A sleep apnea detection method based on a double attention mechanism is characterized by comprising the following steps:
acquiring an original electrocardiosignal, and performing data preprocessing on the original electrocardiosignal to obtain an RRI and an Rpeak;
performing feature extraction on the RRI and the Rpeak through a one-dimensional convolution neural network module to obtain a first feature and a second feature;
performing weight operation processing on the first feature and the second feature through a first attention module to obtain a third feature;
acquiring a fourth characteristic by performing context dependency relationship capture processing on the third characteristic through a bidirectional long-short term memory network module;
performing weight operation processing on the first feature and the fourth feature through a second attention module to obtain a fifth feature;
and classifying the fifth features through a classifier to obtain a classification result.
2. The method for detecting sleep apnea based on dual attention mechanism as claimed in claim 1, wherein said obtaining raw cardiac electrical signals, and performing data preprocessing on said raw cardiac electrical signals to obtain RRI and Rpeak includes:
acquiring an original electrocardiosignal;
extracting a mark segment and a segment adjacent to the mark segment according to the original electrocardiosignal to obtain 5-min-RRI, 5-min-Rpeak and 1-min-RRI-Rpeak;
carrying out length unified processing on the 5-min-RRI, the 5-min-Rpeak and the 1-min-RRI-Rpeak by an interpolation technology;
wherein the 5-min-RRI represents a RRI of 5 minutes in length, the 5-min-Rpeak represents a Rpeak of 5 minutes in length, and the 1-min-RRI-Rpeak represents a RRI and a Rpeak of 1 minute in length.
3. The method of claim 2, wherein the feature extraction of the RRI and the Rpeak by the one-dimensional convolutional neural network module to obtain a first feature and a second feature comprises:
performing feature extraction on the 5-min-RRI and the 5-min-Rpeak through a first one-dimensional convolutional neural network to obtain F'rAnd F'p(ii) a Pairing said F 'in channel dimension'rAnd F'pSplicing to obtain a first characteristic;
performing feature extraction on the 1-min-RRI-Rpeak through a second one-dimensional convolutional neural network to obtain a second feature;
the first one-dimensional convolutional neural network and the second one-dimensional convolutional neural network respectively comprise three one-dimensional convolutional layers and two one-dimensional maximum pooling layers.
4. The method for detecting sleep apnea according to claim 1, wherein said fusing said first feature and said second feature by said first attention module to obtain a third feature comprises:
performing extrusion processing and excitation processing on the first characteristic to obtain a first weight;
and carrying out scale operation on the first weight and the second characteristic to obtain a third characteristic.
5. The sleep apnea detection method based on the dual attention mechanism as recited in claim 1, wherein the obtaining of the fourth feature by performing context dependency capture processing on the third feature through the bidirectional long-short term memory network module comprises:
adjusting, adding and deleting information flow through an LSTM unit, and finishing context dependency relationship capturing processing according to the adjusting, adding and deleting processing;
the bidirectional long-short term memory network module comprises a forward LSTM unit and a backward LSTM unit, and the LSTM unit comprises an input gate, an output gate and a forgetting gate.
6. The sleep apnea detection method based on the dual attention mechanism of claim 1, wherein the obtaining of the fifth feature by performing the weighting operation on the first feature and the fourth feature through the second attention module comprises:
performing extrusion processing and excitation processing on the first features to obtain a second weight;
and carrying out scale operation on the second weight and the fourth feature to obtain a fifth feature.
7. A dual attention mechanism based sleep apnea detection method according to any one of claims 1 to 6, further comprising:
an evaluation index of sleep apnea detection and an apnea hypopnea index are determined.
8. A dual attention mechanism-based sleep apnea detection apparatus, comprising:
the system comprises a first module, a second module and a third module, wherein the first module is used for acquiring an original electrocardiosignal and carrying out data preprocessing on the original electrocardiosignal to obtain RRI and Rpeak;
the second module is used for extracting the features of the RRI and the Rpeak through the one-dimensional convolutional neural network module to obtain a first feature and a second feature;
a third module, configured to perform weight operation on the first feature and the second feature through the first attention module to obtain a third feature;
a fourth module, configured to perform context dependency capture processing on the third feature through a bidirectional long-term and short-term memory network module to obtain a fourth feature;
a fifth module, configured to perform weight operation on the first feature and the fourth feature through the second attention module to obtain a fifth feature;
and the sixth module is used for carrying out classification processing on the fifth characteristic through a classifier to obtain a classification result.
9. An electronic device comprising a processor and a memory;
the memory is used for storing programs;
the processor executing the program realizes the method of any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that the storage medium stores a program which is executed by a processor to implement the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210393579.7A CN114711725A (en) | 2022-04-15 | 2022-04-15 | Sleep apnea detection method and device based on double attention mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210393579.7A CN114711725A (en) | 2022-04-15 | 2022-04-15 | Sleep apnea detection method and device based on double attention mechanism |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114711725A true CN114711725A (en) | 2022-07-08 |
Family
ID=82242946
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210393579.7A Pending CN114711725A (en) | 2022-04-15 | 2022-04-15 | Sleep apnea detection method and device based on double attention mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114711725A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115486833A (en) * | 2022-08-22 | 2022-12-20 | 华南师范大学 | Respiration state detection method and device, computer equipment and storage medium |
| CN116616792A (en) * | 2023-07-21 | 2023-08-22 | 山东大学 | Atrial Fibrillation Detection System Based on Lightweight Design and Feature Fusion |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1151719A2 (en) * | 2000-05-05 | 2001-11-07 | Pacesetter, Inc. | Apparatus for monitoring heart failure via respiratory patterns |
| CN110349676A (en) * | 2019-06-14 | 2019-10-18 | 华南师范大学 | Timing physiological data classification method, device, storage medium and processor |
| WO2020166239A1 (en) * | 2019-02-13 | 2020-08-20 | 国立大学法人京都大学 | Sleep apnea syndrome determination apparatus, sleep apnea syndrome determination method, and sleep apnea syndrome determination program |
| US20200388287A1 (en) * | 2018-11-13 | 2020-12-10 | CurieAI, Inc. | Intelligent health monitoring |
| CN112426147A (en) * | 2020-10-21 | 2021-03-02 | 华南师范大学 | Sleep respiratory event detection model processing method, system and storage medium |
| CN113243890A (en) * | 2021-05-10 | 2021-08-13 | 清华大学深圳国际研究生院 | Sleep apnea syndrome recognition device |
| KR102298119B1 (en) * | 2020-12-10 | 2021-09-03 | 주식회사 에이슬립 | Method and computer device for providing analytical information related to sleep |
| WO2022031038A1 (en) * | 2020-08-04 | 2022-02-10 | 주식회사 에이슬립 | Computing device for predicting sleep state on basis of data measured in sleep environment of user |
| CN114067435A (en) * | 2021-11-15 | 2022-02-18 | 山东大学 | Sleep behavior detection method and system based on pseudo-3D convolutional network and attention mechanism |
-
2022
- 2022-04-15 CN CN202210393579.7A patent/CN114711725A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1151719A2 (en) * | 2000-05-05 | 2001-11-07 | Pacesetter, Inc. | Apparatus for monitoring heart failure via respiratory patterns |
| US20200388287A1 (en) * | 2018-11-13 | 2020-12-10 | CurieAI, Inc. | Intelligent health monitoring |
| WO2020166239A1 (en) * | 2019-02-13 | 2020-08-20 | 国立大学法人京都大学 | Sleep apnea syndrome determination apparatus, sleep apnea syndrome determination method, and sleep apnea syndrome determination program |
| CN110349676A (en) * | 2019-06-14 | 2019-10-18 | 华南师范大学 | Timing physiological data classification method, device, storage medium and processor |
| WO2022031038A1 (en) * | 2020-08-04 | 2022-02-10 | 주식회사 에이슬립 | Computing device for predicting sleep state on basis of data measured in sleep environment of user |
| CN112426147A (en) * | 2020-10-21 | 2021-03-02 | 华南师范大学 | Sleep respiratory event detection model processing method, system and storage medium |
| KR102298119B1 (en) * | 2020-12-10 | 2021-09-03 | 주식회사 에이슬립 | Method and computer device for providing analytical information related to sleep |
| CN113243890A (en) * | 2021-05-10 | 2021-08-13 | 清华大学深圳国际研究生院 | Sleep apnea syndrome recognition device |
| CN114067435A (en) * | 2021-11-15 | 2022-02-18 | 山东大学 | Sleep behavior detection method and system based on pseudo-3D convolutional network and attention mechanism |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115486833A (en) * | 2022-08-22 | 2022-12-20 | 华南师范大学 | Respiration state detection method and device, computer equipment and storage medium |
| CN116616792A (en) * | 2023-07-21 | 2023-08-22 | 山东大学 | Atrial Fibrillation Detection System Based on Lightweight Design and Feature Fusion |
| CN116616792B (en) * | 2023-07-21 | 2023-12-01 | 山东大学 | Atrial fibrillation detection system based on lightweight design and feature fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | A method to detect sleep apnea based on deep neural network and hidden Markov model using single-lead ECG signal | |
| CN111449645B (en) | Intelligent classification and identification method for electrocardiogram and heartbeat | |
| Taji et al. | False alarm reduction in atrial fibrillation detection using deep belief networks | |
| CN114711725A (en) | Sleep apnea detection method and device based on double attention mechanism | |
| Azimi et al. | Machine learning-based automatic detection of central sleep apnea events from a pressure sensitive mat | |
| CN113303814B (en) | Single-channel ear electroencephalogram automatic sleep staging method based on deep transfer learning | |
| CN114391807B (en) | Sleep breathing disorder analysis method, device, equipment and readable medium | |
| Roy et al. | On-device reliability assessment and prediction of missing photoplethysmographic data using deep neural networks | |
| CN116098602B (en) | Non-contact sleep respiration monitoring method and device based on IR-UWB radar | |
| Olsen et al. | Automatic, electrocardiographic-based detection of autonomic arousals and their association with cortical arousals, leg movements, and respiratory events in sleep | |
| Dhyani et al. | Arrhythmia disease classification utilizing ResRNN | |
| CN113925459A (en) | Sleep staging method based on electroencephalogram feature fusion | |
| Shen et al. | Multitask residual shrinkage convolutional neural network for sleep apnea detection based on wearable bracelet photoplethysmography | |
| Cao et al. | Multi-task feature fusion network for Obstructive Sleep Apnea detection using single-lead ECG signal | |
| CN114267444A (en) | Method for detecting obstructive apnea and night frontal epilepsy by using sleep structure | |
| Paul et al. | ECG and SpO2 signal-based real-time sleep apnea detection using feed-forward artificial neural network | |
| Mendonça et al. | A method for sleep quality analysis based on CNN ensemble with implementation in a portable wireless device | |
| Kumar et al. | CACHET-CADB: A contextualized ambulatory electrocardiography arrhythmia dataset | |
| CN116807405A (en) | Sleep state and sleep disease detection system based on human body pressure distribution image | |
| Feng et al. | Machine learning-aided detection of heart failure (LVEF≤ 49%) by using ballistocardiography and respiratory effort signals | |
| Chen et al. | Contactless screening for sleep apnea with breathing vibration signals based on modified U-Net | |
| de Melo | A brief review on electrocardiogram analysis and classification techniques with machine learning approaches | |
| Jothi et al. | Deep Learning Based Obstructive Sleep Apnea Detection for e-health Applications | |
| Sathish et al. | Predicting Sleeping Quality Using Convolutional Neural Networks | |
| Wang et al. | A high precision real-time premature ventricular contraction assessment method based on the complex feature set |
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 |