Key result
A time-frequency and convolutional neural network-based approach using Pseudo Wigner-Ville Distribution detected five ECG abnormalities with an accuracy of 98.9%.
Why the study?
Accurate identification of cardiac pathological events is critical during ECG assessment, which currently relies on manual analysis by cardiologists to detect abnormalities.
Does a Pseudo Wigner-Ville Distribution and Convolutional Neural Network-based approach accurately identify and classify abnormalities in ECG signals?
Does a Pseudo Wigner-Ville Distribution and Convolutional Neural Network-based approach accurately identify and classify abnormalities in ECG signals?
A novel machine learning approach combining Pseudo Wigner-Ville Distribution and Convolutional Neural Networks achieved 98.9% accuracy in classifying five types of ECG abnormalities.
Accurate identification of cardiac pathological events is the most important part of electrocardiogram (ECG) signal assessment. Cardiologists manually analyze data to detect any abnormality in the cardiac signal. Various abnormalities can be found in the cardiac signal such as ventricular tachyarrhythmia, malignant ventricular ectopy, congestive heart failure, intracardiac atrial fibrillation and many others. Here, we propose a time-frequency (TF) and convolution neural network (CNN) based approach for identification and classification of abnormalities in ECG. The proposed methodology applies Pseudo Wigner-Ville Distribution (PWVD) on non-stationary ECG signals to obtain high resolution images and defining the signal in both time and frequency domains. The images train the multi-layer Convolution Neural Network (CNN) for classification. On testing, signatures of five abnormalities have been detected in ECG signals with high accuracy of 98.9%. The technique will facilitate the doctors to accurately diagnose and classify disease symptoms of cardiac patients.
No takes yet. Share an insight, caveat, or question.
Zulfiqar et al. (2019) studied ECG abnormalities. Pseudo Wigner-Ville Distribution (PWVD) and Convolutional Neural Network (CNN) was evaluated on Detection and classification of five ECG abnormalities. A time-frequency and convolutional neural network-based approach using Pseudo Wigner-Ville Distribution detected five ECG abnormalities with an accuracy of 98.9%.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: