Abstract:Advanced deep neural networks, when trained on extensive datasets, can outperform cardiologists in diagnosing cardiac arrhythmias. However, the availability of large-scale training data is often impractical. This study explores the use of transfer learning to identify and classify three ECG patterns. It applies knowledge gained from 2D image classification tasks to the domain of 1D time-series ECG signal classification. The research leverages various deep learning models to classify continuous wavelet transform (2D representations) of ECG signals. The effectiveness of these transferred deep learning models in classifying ECG time-series data is then evaluated.

What problem does this paper attempt to address?

This paper discusses how to use 2D transfer learning to solve the problem of electrocardiogram (ECG) classification. In the absence of a large amount of training data, advanced deep neural networks may outperform cardiac experts in diagnosing cardiac rhythm abnormalities. The study uses continuous wavelet transform to transform 1D time-series ECG signals into 2D images, and then applies pre-trained deep learning models for classification. In this way, the researchers evaluated the effectiveness of these deep learning models in identifying ECG time-series data. Specifically, the paper uses seven different deep learning architectures, ranging from 41 layers to 708 layers, and selects scalograms (wavelet transform images) that preserve more detailed features instead of other transformation methods. The paper also emphasizes the transfer learning potential of pre-trained ImageNet models in time-series signal classification, especially for ECG signals. In the experiments, the study used a dataset containing 162 ECG records with different cardiac conditions. The results showed that the performance of the deep networks improves as the depth of the architecture increases, but it decreases after a certain point. The best performance was achieved by ResNet-50, with an accuracy of 94.12%. This indicates that selecting a suitable pre-trained architecture for time-series analysis is crucial for medical image classification tasks, especially for ECG signal classification. In conclusion, the paper aims to address the problem of data scarcity in ECG classification through transfer learning, proposing an effective method and emphasizing the importance of selecting appropriate deep learning models. This research not only contributes to ECG analysis but also has potential applications in the detection of other cardiovascular diseases.

2D Transfer Learning for ECG Classification using Continuous Wavelet Transform

Deep Arrhythmia Classification Based on SENet and Lightweight Context Transform

ECG Heartbeat Classification Using Deep Transfer Learning with Convolutional Neural Network and STFT Technique

Automatic classification of electrocardiogram signals based on transfer learning and continuous wavelet transform

TLFT: Transfer Learning and Fourier Transform for ECG Classification

DeepECG: Image-based Electrocardiogram Interpretation with Deep Convolutional Neural Networks

Enhancing ECG classification with continuous wavelet transform and multi-branch transformer

Multi-label Classification of Abnormalities in 12-Lead ECG Using Deep Learning.

Arrhythmia Detection from Electrocardiogram Signal Data Based on Wavelet Transform and Deep Reinforcement Learning

2D-wavelet encoded deep CNN for image-based ECG classification

Automatic ECG Classification Using Continuous Wavelet Transform and Convolutional Neural Network

Classification of ECG signals using deep neural networks

A Comparison of 1-D and 2-D Deep Convolutional Neural Networks in ECG Classification

Deep learning research on clinical electrocardiogram analysis

Ensemble Deep Learning Approach for ECG-Based Cardiac Disease Detection: Signal and Image Analysis

A deep learning approach to cardiovascular disease classification using empirical mode decomposition for ECG feature extraction

ELECTROCARDIOGRAM SIGNAL CLASSIFICATION ALGORITHM BASED ON THE CONTINUOUS WAVELET TRANSFORM AND GOOGLENET IN AN INTERNET OF THINGS CONTEXT

Advanced integration of 2DCNN-GRU model for accurate identification of shockable life-threatening cardiac arrhythmias: a deep learning approach

Classification of Arrhythmia by Using Deep Learning with 2-D ECG Spectral Image Representation

Classification of ECG signals for detection of arrhythmia and congestive heart failure based on continuous wavelet transform and deep neural networks

Transfer Learning in ECG Diagnosis: Is It Effective?