Geometric Deep Learning for Topologically Consistent Modeling of Cortical Anatomy
MCML Authors
Fabian Bongratz
Dr.
Abstract
Fabian Bongratz
Dr.
Abstract
Modeling the cerebral cortex is essential for computing and mapping biomarkers that help study neurological diseases, but its complex folded structure is difficult to capture with conventional image analysis. This dissertation presents new surface-based deep learning methods that overcome this limitation. Specifically, new methods for surface reconstruction, prediction, and generative modeling with applications to neuroimaging are presented.
BibTeXKey: Bon26