CN112971799A - Non-stimulation fetal heart monitoring classification method based on machine learning - Google Patents

Non-stimulation fetal heart monitoring classification method based on machine learning Download PDF

Info

Publication number
CN112971799A
CN112971799A CN202110169914.0A CN202110169914A CN112971799A CN 112971799 A CN112971799 A CN 112971799A CN 202110169914 A CN202110169914 A CN 202110169914A CN 112971799 A CN112971799 A CN 112971799A
Authority
CN
China
Prior art keywords
fetal
fetal heart
data
heart rate
heart monitoring
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
Application number
CN202110169914.0A
Other languages
Chinese (zh)
Inventor
唐晓英
张钊
李广飞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Institute of Technology BIT
Original Assignee
Beijing Institute of Technology BIT
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Institute of Technology BIT filed Critical Beijing Institute of Technology BIT
Priority to CN202110169914.0A priority Critical patent/CN112971799A/en
Publication of CN112971799A publication Critical patent/CN112971799A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • 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

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Physics & Mathematics (AREA)
  • Signal Processing (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Surgery (AREA)
  • Physiology (AREA)
  • Biophysics (AREA)
  • Pathology (AREA)
  • Biomedical Technology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Psychiatry (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Evolutionary Computation (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

The invention relates to the field of electronic fetal heart monitoring, in particular to a non-stimulation fetal heart monitoring classification method based on machine learning. On one hand, a fetal electrocardiosignal feature construction method based on a neural network is provided, and on the other hand, fetal heart monitoring is classified by machine learning through an integrated feature index set. The characteristic construction method based on the neural network adopts blind source separation to obtain fetal electrocardiosignals, and uses RNN and CNN to obtain characteristic vectors of the fetal electrocardiosignals; the fetal heart monitoring is classified by machine learning through the integrated characteristic index set, and nonlinear characteristics such as approximate entropy and time domain characteristics such as sleep-wake period are extracted by utilizing effective fetal heart rate; the classification identification module classifies the fetal heart monitoring type by using an integrated characteristic classifier to acquire a fetal NST type. The method provided by the invention can classify the fetal NST types of 24-hour fetal heart monitoring data in real time by using a machine learning method, the classification process is objective, and support is provided for clinical decision.

Description

Non-stimulation fetal heart monitoring classification method based on machine learning
Technical Field
The invention relates to the field of electronic fetal heart monitoring, in particular to a non-stimulation fetal heart monitoring classification method based on machine learning.
Background
The electronic fetal heart monitoring is a fetal intrauterine monitoring technology which is most widely applied clinically, mainly aims at evaluating the development state of a fetus in an uterus, timely finding and treating the conditions of fetal hypoxia and the like, and is beneficial to reducing the morbidity and mortality of perinatal infants.
The research focus of the electronic fetal heart monitoring theory and technology is a pregnant woman abdominal electrode monitoring method, which is to place an electrode on the surface of the pregnant woman abdomen for signal acquisition and then extract a pure fetal electrocardiosignal through a related hardware and software separation algorithm.
Currently, most of prenatal fetal heart monitoring in clinic uses a Non-Stress Test (NST) to evaluate the intrauterine reserve capacity of a fetus, the evaluation process is that a doctor observes a fetal heart rate curve by naked eyes, and the fetal heart rate is an important parameter of fetal heart monitoring and is respectively regulated by the rhythm of the heart, a sympathetic-parasympathetic nervous system and hemodynamics. The NST type of the fetus is classified according to the range of characteristic parameters such as baseline, acceleration, deceleration, heart rate variability, etc. This approach is inefficient and highly subjective, and multiple physicians may have different NST classifications for the same fetus.
Doctors are in a cautious mental state, normal types are easily classified into suspicious types, so that the psychological pressure of pregnant women is increased to a certain extent, and certain medical resources are wasted due to multiple fetal heart monitoring. Therefore, an objective method for automatically extracting fetal electrocardiosignal characteristics and fetal heart rate parameter characteristics by a computer and classifying the characteristics is needed to assist a doctor in making a decision, reduce medical resource waste, reduce errors of visual observation and improve evaluation efficiency.
Disclosure of Invention
The embodiment of the invention aims to provide a non-stimulation fetal heart monitoring classification method based on machine learning.
In a first aspect, the embodiment of the invention provides a method for constructing fetal electrocardiosignal characteristics based on a neural network,
obtaining a fetal electrocardiosignal by adopting a time-frequency blind source separation algorithm (TFBSS), and extracting a feature vector of the fetal electrocardiosignal by utilizing but not limited to a Recurrent Neural Network (RNN) and a Convolutional Neural Network (CNN);
maternal signals were acquired using the abdominal electrode, each signal containing three types of signals: separating the maternal electrocardiosignals, the fetal electrocardiosignals and the additional noise by a time-frequency blind source separation algorithm to obtain the fetal electrocardiosignals; according to the electronic fetal heart monitoring three-level evaluation system, a clinical obstetrician carries out grouping marking on fetal types and corresponding fetal electrocardiosignals, a time sequence is sequentially input by utilizing an RNN (radio network) model, and output characteristic vectors are obtained through back propagation and gradient descent algorithm learning; obtaining a time-frequency diagram of a sample signal by using Hilbert-Huang transform (HHT), and obtaining an output characteristic vector by using the time-frequency diagram as input and using a convolution neural network through a back propagation and gradient descent algorithm; and finally, combining the output feature vectors of the RNN and the CNN to construct a feature index set.
In a second aspect, the embodiments of the present invention classify fetal heart monitoring by machine learning through an integrated feature index set,
and (3) identifying the position of the fetal electrocardiosignal R wave by adopting an improved Pan-Tomkins algorithm, and calculating the fetal heart rate. The method comprises the steps of extracting fetal intrauterine characteristic parameters such as a fetal heart rate baseline, acceleration, variation, a wake-sleep cycle and nonlinear parameters through a separated fetal heart rate and a calculation method, analyzing the mean value and the range of each characteristic parameter of the fetal heart rate under continuous long-term monitoring by using a statistical method, constructing a characteristic index set by using the extracted characteristic indexes with significant differences, finally integrating the characteristic index set of the fetal heart rate into a fetal electrocardiosignal characteristic index set, and classifying the fetal intrauterine state by using a machine learning method.
Drawings
In order to more clearly illustrate the embodiments of the present invention and to enable those skilled in the art to better understand the technical solutions of the present disclosure, reference will now be made to the accompanying drawings used in the embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention.
FIG. 1 is a flow chart of a principal embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The method provided by the invention is mainly used for automatic classification of non-stimulation fetal heart monitoring, and the method comprises two parts: 1. a fetal electrocardiosignal characteristic construction method based on a neural network; 2. and classifying the fetal heart monitoring by machine learning through the integrated characteristic index set.
1. Fetus electrocardiosignal characteristic construction method based on neural network
The observed signal collected by the electrode is UiThe source signal is DiThe noise signal is etaiThe mixing matrix is A and has Ui＝ADi+ηiWherein U isiThe signal is an instantaneous linear mixed signal containing additive noise, and meets the basic requirements of a blind source separation analysis method. The extraction of fetal electrocardiosignals, i.e. the purpose of blind source separation, is to obtain a source signal DiFrom the observation signal U, without the mixing matrix A being knowniIn which D is separatediBy finding a separation matrix B such that
Figure BDA0002934116940000021
Figure BDA0002934116940000022
I.e. the source signal DiTo estimate the optimum.
Figure BDA0002934116940000023
The data in the channel is the fetal electrocardiosignals obtained by separation. Setting the components corresponding to the maternal electrocardiosignals and the noise to zero, reconstructing the rest components by the mixed matrix and back-projecting the rest components to each electrode, thereby obtaining the estimation of the signals of the fetal electrocardiosignals at the body surface electrodes
Figure BDA0002934116940000024
The signal can be approximately considered to contain only fetal cardiac signal sources and noise. Let the input signal be
Figure BDA0002934116940000025
Adopting a time-frequency blind source separation methodAnd extracting the fetal electrocardiosignals, and selecting one path of the output signals as the separated fetal electrocardiosignals.
The method comprises the steps of segmenting a fetal electrocardiosignal data set, adopting 2min as sample duration, increasing data volume through sliding window operation, adopting a bidirectional circulation neural network, using a Gated Current Unit (GRU) in a circulation Unit, adopting a sigmoid function as an activation function, setting neuron in a hidden layer as M, setting a full connection layer as N, adopting a tanh function as an output activation function, adopting cross entropy as a loss function, and obtaining a characteristic vector of 1 multiplied by N through sample data training.
The method comprises the steps of adopting 2min as sample duration, increasing data volume through sliding window operation, obtaining a time-frequency graph of signals through Hilbert-Huang transformation, selecting but not limited to ResNet structures of convolutional neural networks, enabling a convolutional layer activation function to be a relu function, setting a full connection layer to be P, enabling a loss function to be cross entropy, and obtaining a feature vector of 1 xP through training sample data.
2. Classification of fetal heart monitoring by machine learning through an integrated set of feature indices
Mainly comprises three modules (1), namely a fetal heart rate preprocessing module; (2) a fetal heart rate feature extraction module; (3) and a feature integration and classification identification module.
(1) Fetal heart rate preprocessing module
After obtaining the fetal electrocardiosignals through blind source signal separation, firstly, judging the validity, dividing the fetal electrocardiosignals into sequences with the time length of 1 minute, and regarding each minute of sequence signals, when the data loss rate is less than 40%, considering the signals as valid signals. For instruments based on the principle of transabdominal electrocardiosignals, the fetal electrocardiosignals are difficult to extract from maternal and fetal mixed electrocardiosignals due to poor contact between electrode plates and skin.
And for effective fetal electrocardiosignal data, an improved Pan-Tomkins algorithm is adopted to identify the R wave position of the fetal electrocardiosignal and calculate the fetal heart rate.
After the fetal heart rate is obtained, the baseline is extracted according to clinical definition. A baseline value is extracted every 10 minutes, i.e. a window of 10 minutes is followed by a step of 1 minute, and each minute corresponds to a baseline value. Then, in order to ensure that the baseline is a curve with a sampling rate such as the fetal heart rate, interpolation needs to be carried out on the baseline value sequence, and the interpolation method can be selected from, but is not limited to, Lagrange interpolation, Newton interpolation, Hermite interpolation or spline interpolation.
For valid fetal heart rate data, fetal heart rate variability is extracted. The fetal heart rate variation is the difference value between the maximum value and the minimum value of the fetal heart rate in a fixed time. The length of the fixed time period can be freely selected and is generally within 3-10 seconds.
Effective fetal heart rate data, a fetal heart rate baseline sequence and a fetal heart rate variation sequence can be obtained through the preprocessing module.
(2) Fetal heart rate feature extraction module
The module has the main functions of extracting time domain and nonlinear characteristics by using fetal heart rate data and extracting nonlinear characteristics by using a fetal heart rate baseline sequence and a variant sequence.
The time domain characteristics mainly comprise a baseline mean value, a baseline median, a baseline percentile, a variation mean value, a variation median, a variation percentile, an acceleration duration ratio, a unit time acceleration frequency, a unit time acceleration line area, a deceleration duration ratio, a micro variation ratio, a medium variation ratio, a remarkable variation ratio, a unit time quiet sleep frequency and a unit time quiet sleep duration ratio.
The nonlinear characteristics mainly comprise fetal heart rate approximate entropy, sample entropy, association dimension and complexity; approximate entropy of fetal heart rate variability, sample entropy, association dimension, complexity, and approximate entropy of fetal heart rate baseline sequence, sample entropy, association dimension, and complexity.
(3) Feature integration and classification recognition module
The module has the main function of classifying the NST types of the fetuses by using the extracted integrated characteristic parameters.
Non-integrated features refer to features extracted through fetal heart rate, and integrated features refer to all features extracted through fetal electrocardiosignals and fetal heart rate.
Firstly, the characteristics are normalized, namely the fetal electrocardiosignal characteristics and the fetal heart rate characteristic parameters are scaled according to proportion, if the characteristic parameters have positive numbers and negative numbers at the same time, the characteristic parameters are scaled to the value of < -1 >, and if the characteristic parameters have the positive numbers, the characteristic parameters are scaled to the value of 0, 1. The normalization method can be selected from, but is not limited to, the following methods: range transform, 0-mean normalization, linear scale change, etc.
And secondly, classifying the fetal NST test by using the data after the normalization processing as the input of a classifier, wherein the classifier can be selected from but not limited to a support vector machine, a neural network and the like. In the following, taking the support vector machine as an example, the kernel function used is a radial basis kernel function. A specific percentage of individuals is selected as a training set, wherein 90% is taken as an example, the rest 10% of data is taken as a test set, and the training set and the test set cannot have reusable individual data.
For training set individuals, data are trained by using a 5-time cross validation method, and an appropriate error penalty coefficient C and a radial width sigma are found to obtain a classification model with the highest accuracy.
And for the test set individual, inputting the extracted characteristic parameters into the established classifier to obtain the NST type of the test set individual. The classification effect of the classifier on the test set is as shown in the following table, the classification effect of the integrated features is superior to that of the non-integrated features, and the effectiveness of the method is proved.
Actual suspect group Actual Normal group
Predicting suspect groups 19 2
Predicting normal group 4 18
TABLE 1 non-Integrated feature Classification results
Actual suspect group Actual Normal group
Predicting suspect groups 21 1
Predicting normal group 2 19
TABLE 2 Integrated feature Classification results
Figure BDA0002934116940000031
TABLE 3 comparison of technical indices

Claims (2)

1. A fetal electrocardiosignal characteristic construction method based on a neural network is characterized by comprising the following steps:
(1) segmenting a fetal electrocardiosignal data set, adopting 2min as sample duration, increasing data volume through sliding window operation, adopting a bidirectional circulation neural network, using a Gated Current Unit (GRU) as a circulation Unit, adopting a sigmoid function as an activation function, setting neuron in a hidden layer as M, setting a full connection layer as N, adopting a tanh function as an output activation function, adopting cross entropy as a loss function, and obtaining a characteristic vector of 1 multiplied by N through training sample data;
(2) the method comprises the steps of adopting 2min as sample duration, increasing data volume through sliding window operation, obtaining a time-frequency graph of signals through Hilbert-Huang transformation, selecting but not limited to ResNet structures of convolutional neural networks, enabling a convolutional layer activation function to be a relu function, setting a full connection layer to be P, enabling a loss function to be cross entropy, and obtaining a feature vector of 1 xP through training sample data.
2. A method for classifying fetal heart monitoring by machine learning based on an integrated characteristic index set is characterized in that:
(1) after obtaining fetal electrocardiosignals through blind source signal separation, firstly, judging validity, dividing the fetal electrocardiosignals into sequences with the time length of 1 minute, regarding each minute of sequence signals, when the data loss rate is less than 40%, regarding the signals of the section as valid signals, regarding valid fetal electrocardiosignal data, adopting an improved Pan-Tomkins algorithm to identify the R wave position of the fetal electrocardiosignals, and calculating the fetal heart rate;
(2) extracting time domain and nonlinear characteristics through fetal heart rate, wherein the time domain characteristics mainly comprise a baseline mean value, a baseline median, a baseline percentile, a variation mean value, a variation median, a variation percentile, an acceleration duration ratio, unit time acceleration times, unit time offline area of acceleration, a deceleration duration ratio, a tiny variation ratio, a medium variation ratio, a significant variation ratio, unit time quiet sleep times and a quiet sleep duration ratio; the nonlinear characteristics mainly comprise fetal heart rate approximate entropy, sample entropy, association dimension, complexity, fetal heart rate variation approximate entropy, sample entropy, association dimension, complexity, fetal heart rate baseline sequence approximate entropy, sample entropy, association dimension, complexity and the like;
(3) normalizing the integrated feature, if the feature parameter has both positive and negative numbers, scaling to [ -1,1], and if the feature parameter has only positive numbers, scaling to [0,1], wherein the normalization method can be selected from, but not limited to, the following methods: range transform, 0-mean normalization, linear scale variation, etc.;
(4) classifying the fetal NST test by using the data after the normalization processing as the input of a classifier, wherein the classifier can be selected from but not limited to a support vector machine, a neural network and the like; taking a support vector machine as an example, the kernel function used is a radial basis kernel function; selecting a specific percentage of individuals as a training set, taking 90% as an example, using the rest 10% of data as a test set, wherein the training set and the test set cannot have reusable individual data, and for the training set individuals, training the data by using a 5-time cross validation method, and finding out an appropriate error penalty coefficient C and a radial width sigma to obtain a classification model with the highest accuracy.
CN202110169914.0A 2021-02-04 2021-02-04 Non-stimulation fetal heart monitoring classification method based on machine learning Pending CN112971799A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110169914.0A CN112971799A (en) 2021-02-04 2021-02-04 Non-stimulation fetal heart monitoring classification method based on machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110169914.0A CN112971799A (en) 2021-02-04 2021-02-04 Non-stimulation fetal heart monitoring classification method based on machine learning

Publications (1)

Publication Number Publication Date
CN112971799A true CN112971799A (en) 2021-06-18

Family

ID=76349038

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110169914.0A Pending CN112971799A (en) 2021-02-04 2021-02-04 Non-stimulation fetal heart monitoring classification method based on machine learning

Country Status (1)

Country Link
CN (1) CN112971799A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113456064A (en) * 2021-07-15 2021-10-01 广州三瑞医疗器械有限公司 Intelligent interpretation method for prenatal fetal heart monitoring signal
CN114159039A (en) * 2021-11-12 2022-03-11 广州三瑞医疗器械有限公司 Intelligent antenatal fetal heart monitoring model
CN114724720A (en) * 2022-06-10 2022-07-08 北京大学第三医院（北京大学第三临床医学院） Prenatal electronic fetal heart monitoring automatic identification system based on deep learning

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101513345A (en) * 2008-12-26 2009-08-26 华南理工大学 Quick blind source separating fetal electrocardioscanner and detection method
CN101972145A (en) * 2010-10-12 2011-02-16 华南理工大学 Fetus electrocardio blind separation method based on relative sparsity of time domain of source signal
CN103800037A (en) * 2014-01-15 2014-05-21 北京春闱科技有限公司 Fetal heart monitoring system, fetal heart monitoring equipment and fetal heart monitoring method
US20140378855A1 (en) * 2013-06-25 2014-12-25 The Research Foundation For The State University Of New York Apparatus and method for feature extraction and classification of fetal heart rate
EP3189776A1 (en) * 2016-01-08 2017-07-12 Koninklijke Philips N.V. An apparatus and method for generating fetal heart rate data
US20180000405A1 (en) * 2016-07-01 2018-01-04 Bloom Technologies NV Systems and methods for health monitoring
CN107693004A (en) * 2017-09-05 2018-02-16 广东工业大学 Fetal ECG extraction and fetal heart frequency recognition methods based on hilbert conversion
CN108742599A (en) * 2018-03-20 2018-11-06 深圳大学 A kind of foetus health early warning system and method
CN109171708A (en) * 2018-10-25 2019-01-11 广东工业大学 One kind can defibrillation rhythm of the heart identification device
CN110269625A (en) * 2019-05-31 2019-09-24 杭州电子科技大学 A kind of electrocardio authentication method and system of novel multiple features fusion
CN110420019A (en) * 2019-07-29 2019-11-08 西安电子科技大学 A kind of depth recurrence heart rate estimation method of ballistocardiography signal
CN110555388A (en) * 2019-08-06 2019-12-10 浙江大学 CNN and LSTM-based method for constructing intracardiac abnormal excitation point positioning model
CN111449644A (en) * 2020-03-19 2020-07-28 复旦大学 Bioelectricity signal classification method based on time-frequency transformation and data enhancement technology

Patent Citations (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101513345A (en) * 2008-12-26 2009-08-26 华南理工大学 Quick blind source separating fetal electrocardioscanner and detection method
CN101972145A (en) * 2010-10-12 2011-02-16 华南理工大学 Fetus electrocardio blind separation method based on relative sparsity of time domain of source signal
US20180078160A1 (en) * 2013-06-25 2018-03-22 The Research Foundation for The State University of New York. Apparatus and method for feature extraction and classification of fetal heart rate
US20140378855A1 (en) * 2013-06-25 2014-12-25 The Research Foundation For The State University Of New York Apparatus and method for feature extraction and classification of fetal heart rate
CN103800037A (en) * 2014-01-15 2014-05-21 北京春闱科技有限公司 Fetal heart monitoring system, fetal heart monitoring equipment and fetal heart monitoring method
EP3189776A1 (en) * 2016-01-08 2017-07-12 Koninklijke Philips N.V. An apparatus and method for generating fetal heart rate data
US20180000405A1 (en) * 2016-07-01 2018-01-04 Bloom Technologies NV Systems and methods for health monitoring
CN107693004A (en) * 2017-09-05 2018-02-16 广东工业大学 Fetal ECG extraction and fetal heart frequency recognition methods based on hilbert conversion
CN108742599A (en) * 2018-03-20 2018-11-06 深圳大学 A kind of foetus health early warning system and method
CN109171708A (en) * 2018-10-25 2019-01-11 广东工业大学 One kind can defibrillation rhythm of the heart identification device
CN110269625A (en) * 2019-05-31 2019-09-24 杭州电子科技大学 A kind of electrocardio authentication method and system of novel multiple features fusion
CN110420019A (en) * 2019-07-29 2019-11-08 西安电子科技大学 A kind of depth recurrence heart rate estimation method of ballistocardiography signal
CN110555388A (en) * 2019-08-06 2019-12-10 浙江大学 CNN and LSTM-based method for constructing intracardiac abnormal excitation point positioning model
CN111449644A (en) * 2020-03-19 2020-07-28 复旦大学 Bioelectricity signal classification method based on time-frequency transformation and data enhancement technology

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
李广飞: "电子胎心监护中胎心率特征提取分析及其应用探究", 《中国优秀硕士学位论文全文数据库医药卫生科技》 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113456064A (en) * 2021-07-15 2021-10-01 广州三瑞医疗器械有限公司 Intelligent interpretation method for prenatal fetal heart monitoring signal
CN113456064B (en) * 2021-07-15 2024-04-16 广州三瑞医疗器械有限公司 Intelligent interpretation method for prenatal fetal heart monitoring signals
CN114159039A (en) * 2021-11-12 2022-03-11 广州三瑞医疗器械有限公司 Intelligent antenatal fetal heart monitoring model
CN114724720A (en) * 2022-06-10 2022-07-08 北京大学第三医院（北京大学第三临床医学院） Prenatal electronic fetal heart monitoring automatic identification system based on deep learning

Similar Documents

Publication Publication Date Title
Jager et al. Characterization and automatic classification of preterm and term uterine records
Wang et al. Detection analysis of epileptic EEG using a novel random forest model combined with grid search optimization
CN112971799A (en) Non-stimulation fetal heart monitoring classification method based on machine learning
Boashash et al. Automatic signal abnormality detection using time-frequency features and machine learning: A newborn EEG seizure case study
Cömert et al. Fetal hypoxia detection based on deep convolutional neural network with transfer learning approach
Fergus et al. Advanced artificial neural network classification for detecting preterm births using EHG records
Fergus et al. Classification of caesarean section and normal vaginal deliveries using foetal heart rate signals and advanced machine learning algorithms
Acharya et al. Automated diagnosis of epileptic EEG using entropies
Zheng et al. A novel hybrid energy fraction and entropy-based approach for systolic heart murmurs identification
Agrawal et al. Cardiotocography analysis for fetal state classification using machine learning algorithms
CN104887224A (en) Epileptic feature extraction and automatic identification method based on electroencephalogram signal
Cömert et al. Evaluation of fetal distress diagnosis during delivery stages based on linear and nonlinear features of fetal heart rate for neural network community
Degbedzui et al. Accurate diagnosis of term–preterm births by spectral analysis of electrohysterography signals
Subramaniam et al. A review of significant researches on prediction of preterm birth using uterine electromyogram signal
Lu et al. Estimation of the foetal heart rate baseline based on singular spectrum analysis and empirical mode decomposition
Boashash et al. Time-frequency signal and image processing of non-stationary signals with application to the classification of newborn EEG abnormalities
Shahrdad et al. Detection of preterm labor by partitioning and clustering the EHG signal
Abel et al. A comprehensive survey on signal processing and machine learning techniques for non-invasive fetal ECG extraction
CN113076878A (en) Physique identification method based on attention mechanism convolution network structure
Ge et al. Detection of pulmonary hypertension associated with congenital heart disease based on time-frequency domain and deep learning features
CN109033990B (en) CNN model heartbeat classification method based on intra-class and inter-class distances
Bhatnagar et al. Classification of cardiotocography data with WEKA
CN113011330B (en) Electroencephalogram signal classification method based on multi-scale neural network and cavity convolution
Liu et al. Multiscale sample entropy analysis of wrist pulse blood flow signal for disease diagnosis
Gnana Rajesh Analysis of MFCC features for EEG signal classification

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20210618