Motion artifact recognition and quantification in coronary CT angiography using convolutional neural networks

Tanja Lossau, Hannes Nickisch, Tobias Wissel, Rolf Bippus, Holger Schmitt, Michael M. Morlock, Michael Grass

Abstract: Highlights•Coronary Motion Forward Artifact model for CT data (CoMoFACT) induces simulated motion by means of the motion compensated filtered back-projection algorithm.•Realistic data for supervised learning generated by the CoMoFACT.•First machine-learning-based measures for coronary motion artifact recognition and quantification.•Higher robustness regarding variations in background intensities compared to state of the art handcrafted measures.•Applicability for image quality assessment in ECG-triggered and ECG-gated CT scans.