Skip to main content
. Author manuscript; available in PMC: 2020 Dec 1.
Published in final edited form as: NMR Biomed. 2020 Feb 20;33(12):e4271. doi: 10.1002/nbm.4271

Figure 4:

Figure 4:

VaNDI acts as an unrolled gradient descent algorithm combining deep learning and nonlinear data-fidelity. In the CNN part of the network, 3D convolutions (green) and nonlinear activation (blue) are learned for each gradient descent step GDt. Moreover, a data-fidelity term (orange) is integrated into the network to ensure agreement with the QSM model.