Automatic Premature Ventricular Contractions Detection for Multi-Lead Electrocardiogram Signal

In this paper, we propose an electrocardiogram (ECG) technique for the automatic detection of Premature Ventricular Contractions (PVC) based on multi-lead signals and on a deep learning architecture which is built using Stacked Denoising Autoencoders (SDAEs) networks. The proposed method consists of...

Full description

Saved in:
Bibliographic Details
Main Authors: Rahhal, Mohamad Mahmoud Al, Ajlan, Naif Al, Bazi, Yakoub, Hichri, Haikel Al, Rabczuk, Timon
Format: Conference Proceeding
Language:English
Subjects:
Electrocardiography
Electrodes
Heart beat
Heart rate variability
Neural networks
Training
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we propose an electrocardiogram (ECG) technique for the automatic detection of Premature Ventricular Contractions (PVC) based on multi-lead signals and on a deep learning architecture which is built using Stacked Denoising Autoencoders (SDAEs) networks. The proposed method consists of two main stages; feature learning and classification. In the first stage, we learn a new feature representation from data using SDAEs. Regarding the classification, we add a softmax regression layer on the top of the resulting hidden representation layer yielding a deep neural network (DNN). The proposed method fuses the results of several ECG leads (up to 12) in order to increase the detection accuracy. In the experiments, we use INCART database to test the proposed DNN multi-lead method. The obtained results are 98.6%, 91.4%, and 97.7% respectively for overall accuracy (OA), average sensitivity (Se), and average positive productivity (Pp).
ISSN:2154-0373
DOI:10.1109/EIT.2018.8500197