← All publications
Image segmentationSupports VenusX Sizing
Segmentation of the aorta in systolic phase from 4D flow MRI: multi-atlas vs. deep learning
Magnetic Resonance Materials in Physics, Biology and Medicine
Question
How do atlas-based and deep learning methods compare for automatic aortic segmentation?
What the paper shows
Deep learning and multi-atlas segmentation were compared on 36 patients, both on the segmentation itself and on the wall pressure computed from each model. Deep learning gave the more accurate segmentation and the pressures closest to those from manual segmentation.
Authors
Marin-Castrillon D. M., Geronzi L., Boucher A., Lin S., Morgant M.-C., Cochet A., Rochette M., Leclerc S., Ambarki K., Jin N., Aho L. S., Lalande A., Bouchot O., Presles B.
