Reconstructing Complex Cardiac Excitation Waves From Incomplete Data Using Echo State Networks and Convolutional Autoencoders

Echo state networks (ESNs) and convolutional autoencoders (CAEs) are applied to solve two data modelling tasks in cardiac dynamics: Recovering excitation patterns from from noisy, blurred or undersampled observations and reconstructing complex electrical excitation waves from mechanical deformation. Both approaches provide satisfying solutions, but CAEs turned out to be superior to ESNs in terms of reconstruction errors.