Resolution translation for GRE using super-resolution network

Motivation: SWI is a widely acquired protocol in clinic and often acquired in an anisotropic resolution (i.e., 0.5×0.5×2 mm3). Whereas QSM, which has been attracting utilities, has been recommended to acquire in 1 mm isotropic resolution. Goal(s): This study is aimed to propose a resolution translation method so that SWI data can be transformed to QSM reconstruction. Approach: A super-resolution network was designed to transform complex GRE data, exploiting high in-plane resolution information while sustaining down-sampling consistency. Results: The proposed method successfully generated high-quality isotropic resolution data for QSM from anisotropic SWI data. Impact: The proposed resolution translation successfully generated high-quality complex GRE images for the isotropic resolution, resulting in high-quality QSM, outperforming the conventional approaches.