Navigating Cardiac Complexity using CNN Autoencoder Solutions for Detecting Abnormal Heart Signals

This study examines the application of Convolutional Neural Network (CNN) Autoencoders in identifying abnormalities in electrocardiogram (ECG) data using the PTB Diagnostic ECG Database. The dataset comprises 14,552 samples categorised into two groups: normal heartbeats and heartbeats affected by ca...

Full description

Saved in:
Bibliographic Details
Published in2024 OPJU International Technology Conference (OTCON) on Smart Computing for Innovation and Advancement in Industry 4.0 pp. 1 - 5
Main Authors Agarwal, Muskan, Rajput, Kapil, Gill, Kanwarpartap Singh, Singh, Vijay
Format Conference Proceeding
LanguageEnglish
Published IEEE 05.06.2024
Subjects
Accuracy
Anomaly detection
Artificial Intelligence
Autoencoder
Convolution
Deep Learning
Electrocardiography
Heart beat
Measurement
Medical services
Model Training
Online AccessGet full text

Cover

More Information
Summary:This study examines the application of Convolutional Neural Network (CNN) Autoencoders in identifying abnormalities in electrocardiogram (ECG) data using the PTB Diagnostic ECG Database. The dataset comprises 14,552 samples categorised into two groups: normal heartbeats and heartbeats affected by cardiac disorders. The utilisation of Transposed Convolution has significantly improved the efficiency of the model. The study emphasises the critical significance of immediately detecting cardiac irregularities and offers a Convolutional Neural Network (CNN) Autoencoder model particularly designed to efficiently encode and decode electrocardiogram (ECG) data, hence aiding in the identification of irregular patterns. The procedure involves building a robust Autoencoder that consists of encoder and decoder components, which are trained to reduce errors in reconstruction. The assessment metrics showcase the model's outstanding accuracy (76.93 \%), precision (55.23 \%), recall \mathbf{(} 89.81 \%), and F1 score (65.40 \%). This work highlights the need of utilising deep learning techniques to detect abnormalities in ECG data at an early stage. This technique has significant promise in strengthening diagnostic abilities and improved patient results.
DOI:10.1109/OTCON60325.2024.10687934