Ventricular ectopic beat detection using a wavelet transform and a convolutional neural network.

08:00 EDT 10th April 2019 | BioPortfolio

Summary of "Ventricular ectopic beat detection using a wavelet transform and a convolutional neural network."

\emph{Objective:} Ventricular contractions in healthy individuals normally follow the contractions of atria to facilitate more efficient pump action and cardiac output. With a ventricular ectopic beat (VEB), volume within the ventricles are pumped to the body's vessels before receiving blood from atria, thus causing inefficient blood circulation. VEBs tend to cause perturbations in the instantaneous heart rate time series, making the analysis of heart rate variability inappropriate around such events, or requiring special treatment (such as signal averaging). Moreover, VEB frequency can be indicative of life-threatening problems. However, VEBs can often mimic artifacts both in morphology and timing. Identification of VEBs is therefore an important unsolved problem. The aim of this study is to introduce a method of wavelet transform in combination with deep learning network for the classification of VEBs. \emph{Approach:} We proposed a method to automatically discriminate VEB beats from other beats and artifacts with the use of wavelet transform of the electrocardiogram (ECG) and a convolutional neural network (CNN). Three types of wavelets (Morlet wavelet, Paul wavelet and Gaussian Derivative) were used to transform segments of single channel (1-D) ECG waveforms to 2-D time-frequency `images'. The 2-D time-frequency images were then passed into a CNN to optimize the convolutional filters and classification. Ten-fold cross validation was used to evaluate the approach on the MIT-BIH arrhythmia database (MIT-BIH). The American Heart Association (AHA) database was then used as an independent dataset to evaluate the trained network. \emph{Main results:} Ten-fold cross validation results on MIT-BIH showed that the proposed algorithm with Paul wavelet achieved an overall F1 score of 84.94% and accuracy of 97.96% on out of sample validation. Independent test on AHA resulted in an F1 score of 84.96% and accuracy of 97.36%. \emph{Significance:} The trained network possessed exceptional transferability across databases and generalization to unseen data.


Journal Details

This article was published in the following journal.

Name: Physiological measurement
ISSN: 1361-6579


DeepDyve research library

PubMed Articles [8829 Associated PubMed Articles listed on BioPortfolio]

Wavelet analysis of higher harmonics in tapping mode atomic force microscopy.

Higher harmonics have been widely used to characterize nanomechanical properties of the sample surface in tapping mode atomic force microscopy. They are usually analyzed by the Fourier transform metho...

Onset detection of surface diaphragmatic electromyography based on sample entropy and individualized threshold.

The diaphragm is the main respiratory muscle in the body. The onset detection of the surface diaphragmatic electromyography (sEMGdi) can be used in the respiratory rehabilitation training of the hemip...

Beat-to-beat detection of aortic valve opening in heartware left ventricular assist device patients.

cfLVADs result in a significant reduction in aortic valve (AV) opening, which have been associated with several complications. Reliable monitoring of AV opening is needed to determine whether pump spe...

Research on an adaptive filter for the Mie lidar signal.

The statistical properties of the noise in the Mie lidar signal are analyzed by the statistical hypotheses testing method. Based on this, an adaptive filter is proposed to eliminate the noise. The lea...

Denoising Brain Images with the Aid of Discrete Wavelet Transform and Monarch Butterfly Optimization with Different Noises.

The aim of this paper is to denoise the high noise density image proficiently with minimal computation cost using various techniques. This paper proposes an approach for image denoising based on Discr...

Clinical Trials [3439 Associated Clinical Trials listed on BioPortfolio]

FLuid Responsiveness Prediction Using EXtra Systoles

In this study, the investigators propose to investigate a novel technique for fluid responsiveness prediction. It is based on the occurrence of an extra systole, which induces a preload va...

Fluid Responsiveness Prediction Using Extra Systoles

Fluids are generally administered to patients in order to improve circulation. However, fluids do not always improve circulation and fluids have side effects. Unfortunately, it is difficul...

Evaluation of the Axonemal Dynein Heavy Chain 5 and Creatine Kinase Concentration in Cervical Fluid for Early Detection of the Ectopic Pregnancy

Patients who have doubt for ectopic pregnancy will be enrolled for our study. These patients will be determined with using irregular increased human chorionic gonadotropin (beta-HCG) resul...

PROTECT-PACE STUDY - The Protection of Left Ventricular Function During Right Ventricular Pacing

This study will be done in patients who require the implantation of a cardiac pacemaker (an electronic device that controls the heartbeat) for complete heart block (a heart rhythm abnormal...

Ectopic Pregnancy Biomarkers

In an attempt to simplify the diagnosis of ectopic pregnancy, the researchers will investigate the presence of specific molecular markers of ectopic pregnancy. The quantification of these...

Medical and Biotech [MESH] Definitions

Signal and data processing method that uses decomposition of wavelets to approximate, estimate, or compress signals with finite time and frequency domains. It represents a signal or data in terms of a fast decaying wavelet series from the original prototype wavelet, called the mother wavelet. This mathematical algorithm has been adopted widely in biomedical disciplines for data and signal processing in noise removal and audio/image compression (e.g., EEG and MRI).

A type of automatic, not reentrant, ectopic ventricular rhythm with episodes lasting from a few seconds to a minute which usually occurs in patients with acute myocardial infarction or with DIGITALIS toxicity. The ventricular rate is faster than normal but slower than tachycardia, with an upper limit of 100 -120 beats per minute. Suppressive therapy is rarely necessary.

A potentially lethal cardiac arrhythmia characterized by an extremely rapid, hemodynamically unstable ventricular tachycardia (150-300 beats/min) with a large oscillating sine-wave appearance. If untreated, ventricular flutter typically progresses to VENTRICULAR FIBRILLATION.

A group of cardiac arrhythmias in which the cardiac contractions are not initiated at the SINOATRIAL NODE. They include both atrial and ventricular premature beats, and are also known as extra or ectopic heartbeats. Their frequency is increased in heart diseases.

Detection of drugs that have been abused, overused, or misused, including legal and illegal drugs. Urine screening is the usual method of detection.

Quick Search


DeepDyve research library

Relevant Topics

Cardiology is a specialty of internal medicine.  Cardiac electrophysiology : Study of the electrical properties and conduction diseases of the heart. Echocardiography : The use of ultrasound to study the mechanical function/physics of the h...

Cardiovascular disease (CVD)
Acute Coronary Syndromes (ACS) Blood Cardiovascular Dialysis Hypertension Stent Stroke Vascular Cardiovascular disease (CVD) includes all the diseases of the heart and circulation including coronary heart disease (angina...

Searches Linking to this Article