Improved myelin water fraction mapping with deep neural networks using synthetically generated 3D data

Serge Vasylechko Didenko, Simon K. Warfield, Sila Kurugol, Onur Afacan

Abstract: Highlights•A generative model for synthesis of large scale 3D datasets for myelin mapping.•Combines MR signal decay model, parametric T2 model and spatial generative model.•Targeted for training neural networks that capture spatial and temporal information.•Spatial information helps to regularize solutions for ill posed inverse problems.•MWF estimation from noisy multi-echo T2 data is an ill posed inverse problem.

External IDs:dblp:journals/mia/DidenkoWKA24