Training neural networks with synthetic electrocardiograms

TL;DR

This paper introduces a synthetic data generation method for training neural networks on electrocardiogram signals, achieving comparable or better performance than real data, while enabling privacy and class balance control.

Contribution

The authors develop a domain randomization approach to generate synthetic ECG data that effectively trains neural networks without manual annotations or real data, enhancing privacy and data balance.

Findings

Synthetic data achieves comparable or better performance than real data.

Models trained with synthetic data are robust across different seeds and test sets.

The method enables privacy-preserving and balanced training data generation.

Abstract

We present a method for training neural networks with synthetic electrocardiograms that mimic signals produced by a wearable single lead electrocardiogram monitor. We use domain randomization where the synthetic signal properties such as the waveform shape, RR-intervals and noise are varied for every training example. Models trained with synthetic data are compared to their counterparts trained with real data. Detection of r-waves in electrocardiograms recorded during different physical activities and in atrial fibrillation is used to compare the models. By allowing the randomization to increase beyond what is typically observed in the real-world data the performance is on par or superseding the performance of networks trained with real data. Experiments show robust performance with different seeds and training examples on different test sets without any test set specific tuning. The method makes possible to train neural networks using practically free-to-collect data with accurate labels without the need for manual annotations and it opens up the possibility of extending the use of synthetic data on cardiac disease classification when disease specific a priori information is used in the electrocardiogram generation. Additionally the distribution of data can be controlled eliminating class imbalances that are typically observed in health related data and additionally the generated data is inherently private.