FOD-Net: A deep learning method for fiber orientation distribution angular super resolution

Rui Zeng, Jinglei Lv, He Wang, Luping Zhou, Michael Barnett, Fernando Calamante, Chenyu Wang

Abstract: Highlights•FOD-Net is proposed as an efficient deep learning framework for FOD angular super resolution.•FOD-Net trained and inferenced in the spherical harmonics dimension, and it is less dependent on the diffusion MRI protocol used.•Super-resolved FOD images, high-quality tractograms, and ultimately more reliable structural connectomes can be generated using FOD-Net from common clinical scanners.•State-of-the-art performance tested and validated on the Human Connectome Project dataset and in-house clinical data acquired with clinical GE 3.0T scanner.

External IDs:doi:10.1016/j.media.2022.102431