Deep neural networks for ECG-free cardiac phase and end-diastolic frame detection on coronary angiographies

Open Webpage

Costin Florian Ciusdel, Alexandru Turcea, Andrei Puiu, Lucian Mihai Itu, Lucian Calmac, Emma Weiss, Cornelia Margineanu, Elisabeta Badila, Martin Berger, Thomas Redel, Tiziano Passerini, Mehmet Akif Gülsün, Puneet Sharma

Published: 01 Jan 2020, Last Modified: 04 Nov 2025Comput. Medical Imaging Graph. 2020EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Cardiac phase detection on coronary angiographies is important for the assessment of CAD.•A purely image and deep learning based workflow is introduced for cardiac phase detection.•77,435 coronary angiographies images acquired from 13,081 patients are employed.•Cardiac phase detection has an accuracy of 98.8 %, end diastolic frame detection a precision of 98.4 %.•Detection performance is independent from acquisition angles, heart rate and angiographic view.