Deep learning for cardiac MRI: automatically segmenting left atrium expert human level performance
Radiological Society of North America (RSNA), 103rd Scientific Assembly and Annual Meeting
Abstract
An abstract is not available in the lab archive. View the paper at its source.
An oral presentation at the RSNA 103rd Scientific Assembly and Annual Meeting described a deep-learning approach to segmenting the left atrium and proximal pulmonary veins in cardiac MRI, evaluated using a public challenge dataset.
Summary
This RSNA 2017 presentation described a fully automatic method for segmenting the left atrium and proximal pulmonary veins from cardiac magnetic resonance images. The authors evaluated their approach using data from the STACOM 2013 image segmentation challenge. The RSNA abstract reports Dice and surface-distance metrics and compares the method’s performance with expert human segmentation; these are results from the reported challenge evaluation, not evidence of prospective clinical validation.
Cite this work
Aliasghar Mortazi; Jeremy R. Burt; Ulas Bagci (2017). Deep learning for cardiac MRI: automatically segmenting left atrium expert human level performance. Radiological Society of North America (RSNA), 103rd Scientific Assembly and Annual Meeting https://archive.rsna.org/2017/MagneticResonanceImaging.pdf
BibTeX