Realistic training data and vectorcardiographic improvements of inverse reconstruction

At the Computing in Cardiology conference, we presented two ideas that seem to improve the inverse reconstruction of electrical heart activity. In the first, we propose to use a (computer generated) training set of realistic heart activity as building blocks for reconstructed electrograms at the heart surface. The second idea is to improve inversely reconstructed electrograms by matching their vectorcardiographic characteristics from those observed at the body surface. Find the (award winning) poster below and the paper here, and don’t hesitate to contact us with your ideas and suggestions!

Reference: Matthijs JM Cluitmans, Pietro Bonizzi, Joel MH Karel, Paul GA Volders, Ralf LM Peeters, and Ronald L Westra. Inverse reconstruction of epicardial potentials improves by vectorcardiography and realistic potentials. In Computing in Cardiology, 2013.