Performance Comparison Between Transform-Based Deep Learning Approaches for ECG Signal Classification

• Published in: International Conference on Electrical and Electronics Engineering pp. 439 - 443
• Main Authors: Alboghbaish, Ebrahim, Eleyan, Alaa, Eleyan, Gulden
• Format: Conference Proceeding
• Language: English
• Published: IEEE 22.04.2024
• Subjects: arrhythmia; CNN; Convolutional neural networks; Deep learning; Discrete wavelet transforms; ECG classification; Electrocardiography; Fast Fourier transforms; Feature extraction; Long short term memory; LSTM; Transforms; Vectors
• ISSN: 2694-3646
• DOI: 10.1109/ICEEE62185.2024.10779262

In this paper, we studied and compared the benefits of applying transformation as a preprocessing step to the input of the deep learning algorithms. A comparison was conducted between the discrete wavelet transform (DWT) and fast Fourier transform (FFT) as feature extractors for the ECG signals. The generated feature vectors were fed to a convolutional neural network and long short-term memory (CNN-LSTM) model for classification. Extensive experiments were executed on ECG signals collected from MIT-BIH and BIDMC databases with 3 classes. The obtained results showed the combined power of transformation and deep learning which led to better classification accuracies.