An EMD-Based Approach for Atrial Fibrillation Classification Using Wavelets and Convolutional Neural Network

Background and objectives: Atrial fibrillation (AF) is a common and extremely harmful arrhythmia disease that leads to stroke and death. Several kinds of AF exist as paroxysmal AF, persistent AF, and permanent or chronic AF. Based on electrocardiograms (ECG), wavelet transform (WT), and convolutional neural networks (CNNs), the present study proposes an algorithm to classify between normal subjects and patients suffering from paroxysmal atrial fibrillation (PAF).Methods: We used, from the MIT-BIH ``PAFDB" database, 5-min signals before and during AF. Then, a finite number of intrinsic mode functions (IMFs) were obtained using empirical mode decomposition (EMD). In order to select the most informative IMFs, we propose two criteria: 'variance ratio' and 'percentage of occurrence of IMFs'. The selected IMFs serve as inputs to the continuous wavelet transform (CWT) that extracts features as inputs to our CNN classification model, trained on 80% and tested on 20% of the data.Results: For derivations 0 and 1, respectively, we achieve a maximum accuracy of 98.1% and 98.3% for the 'variance ratio' selection and 98.5% and 98.8% for the "percentage of occurrence of IMFs', compared to 97.5% and 97.6% before the application of these approaches. In addition, the presented approaches reduce the number of IMFs by at least 50%.Conclusion: By combining the variance ratio and percentages of IMFs, the extracted features allow us to classify ECG from healthy patients and patients suffering from AF. The accuracy obtained outperforms some recent studies