Key result
A spectrogram-based 2-D CNN model classifies ECG arrhythmias with 97% accuracy.
Why the study?
Careful analysis of ECG signals is needed for the accurate diagnosis of acute and chronic heart conditions.
Does a 2-D CNN model using spectrograms improve the classification accuracy of arrhythmias from ECG signals?
Does a 2-D CNN model using spectrograms improve the classification accuracy of arrhythmias from ECG signals?
A 2-D CNN model using spectrograms achieves high accuracy (97%) in classifying arrhythmias from ECG signals.
Requires external validation before clinical use; leaves open spectrogram CNN utility in arrhythmia detection.
One of the most commonly used signals in the detection and prediction of cardiovascular diseases is the electrocardiogram (ECG)). The ECG signals can detect rhythmic disturbances in the heart, also known as arrhythmias. For accurate diagnoses of patients acute and chronic heart conditions, careful analysis of ECG signals is needed. In this paper, the classification of ECG signals was done in two classes namely arrhythmia and regular heart rhythm using two-dimensional (2-D) convolutional neural network (CNN) model. Short-time Fourier transform is used to translate one-dimensional ECG time series signals into two-dimensional spectrograms. MIT-BIH arrhythmia dataset is used to test the proposed approach. The average classification test accuracy of 97% and training efficiency of 92.53% is achieved by using this model which is better than conventional methods of classifying arrhythmias. Performance metrices such as sensitivity and specificity etc., indicates significant improvement in showing the efficacy of the proposed CNN model.
No takes yet. Share an insight, caveat, or question.
Kasthuri et al. (2023) studied Cardiac Arrhythmia. 2-D convolutional neural network (CNN) model using spectrograms vs. Conventional methods was evaluated on Classification test accuracy. A 2-D convolutional neural network model using spectrograms classified ECG signals into arrhythmia and regular rhythms with an average test accuracy of 97%.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: