A Transfer-Learning Based Ensemble Architecture for ECG Signal Classification

TL;DR

This paper introduces an ensemble transfer learning approach using modified CNN architectures to classify ECG signals as images, achieving high accuracy and efficiency over traditional methods.

Contribution

It proposes a novel ensemble of transfer learning models with modifications for ECG classification, significantly improving accuracy and reducing runtime.

Findings

Achieved 99.98% accuracy on Physionet dataset

Ensemble method increased accuracy by 6.36% over previous algorithms

Reduced runtime compared to generic CNN architectures

Abstract

Manual interpretation and classification of ECG signals lack both accuracy and reliability. These continuous time-series signals are more effective when represented as an image for CNN-based classification. A continuous Wavelet transform filter is used here to get corresponding images. In achieving the best result generic CNN architectures lack sufficient accuracy and also have a higher run-time. To address this issue, we propose an ensemble method of transfer learning-based models to classify ECG signals. In our work, two modified VGG-16 models and one InceptionResNetV2 model with added feature extracting layers and ImageNet weights are working as the backbone. After ensemble, we report an increase of 6.36% accuracy than previous MLP-based algorithms. After 5-fold cross-validation with the Physionet dataset, our model reaches an accuracy of 99.98%.