Key result
Adding a recurrent model (LSTM) to a convolutional neural network increased its quality index performance by approximately 3% for ECG signal quality assessment, while GAN data augmentation improved CNN performance by up to 2%.
Why the study?
Signal quality assessment is an important procedure to assess if a signal can be preprocessed into the next step.
Enhancing CNNs with recurrent models (LSTM) or GAN-based data augmentation improves the classification of imbalanced wearable ECG signal quality.
Enhances ECG quality classification in computational models; hypothesis-generating, prospective clinical validation required before practice impact.
Signal quality assessment is important procedure to assess if a signal can be preprocessed into next step or not. In this study, we proposed an improving of convolutional neural network with recurrent model and generative adversarial network. The experiment shows that enhancing convolutional neural network by using recurrent model or data augmentation via GAN for wearable ECG signal quality assessment is feasible to try. Adding recurrent model to CNN increase its performance by approximately 3% margin in un-augmented dataset. Generating fake samples from “unacceptable” class can enhance CNN performance up to 2% margin. Combination of recurrent model and GAN still not yet matched found. However, It doesn’t mean that the combination is impossible to do.
No takes yet. Share an insight, caveat, or question.
M. Anwar Ma’sum (2022) studied ECG signal quality assessment (n=1,000). Convolutional Neural Network (CNN) with Recurrent Model (LSTM) and/or GAN data augmentation vs. CNN alone and classical machine learning models was evaluated on Quality index (square root of sensitivity multiplied by specificity). Adding a recurrent model (LSTM) to a convolutional neural network increased its quality index performance by approximately 3% for ECG signal quality assessment, while GAN data augmentation improved CNN performance by up to 2%.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: