Conditional Electrocardiogram Generation Using Hierarchical Variational Autoencoders
This work addresses data scarcity in medical AI for cardiovascular disease diagnosis, offering a synthetic data generation solution that improves model performance, though it is incremental as it builds on existing VAE approaches.
The paper tackles the problem of insufficient training data for machine learning models in cardiovascular disease diagnostics by proposing a conditional hierarchical variational autoencoder (cNVAE-ECG) for generating synthetic 12-lead ECG signals with multiple pathologies, achieving up to a 2% increase in AUROC in downstream tasks compared to GAN-based methods.
Cardiovascular diseases (CVDs) are disorders impacting the heart and circulatory system. These disorders are the foremost and continuously escalating cause of mortality worldwide. One of the main tasks when working with CVDs is analyzing and identifying pathologies on a 12-lead electrocardiogram (ECG) with a standard 10-second duration. Using machine learning (ML) in automatic ECG analysis increases CVD diagnostics' availability, speed, and accuracy. However, the most significant difficulty in developing ML models is obtaining a sufficient training dataset. Due to the limitations of medical data usage, such as expensiveness, errors, the ambiguity of labels, imbalance of classes, and privacy issues, utilizing synthetic samples depending on specific pathologies bypasses these restrictions and improves algorithm quality. Existing solutions for the conditional generation of ECG signals are mainly built on Generative Adversarial Networks (GANs), and only a few papers consider the architectures based on Variational Autoencoders (VAEs), showing comparable results in recent works. This paper proposes the publicly available conditional Nouveau VAE model for ECG signal generation (cNVAE-ECG), which produces high-resolution ECGs with multiple pathologies. We provide an extensive comparison of the proposed model on various practical downstream tasks, including transfer learning scenarios showing an area under the receiver operating characteristic (AUROC) increase up to 2% surpassing GAN-like competitors.