Fig. 11 | Scientific Reports

Fig. 11

From: Complex-valued neural networks to speed-up MR thermometry during hyperthermia using Fourier PD and PDUNet

Fig. 11

The reconstruction quality achieved by the complex-valued models - Fourier-PDNet and Fourier-PDUNet, compared against the zero-padded k-space (denoted as undersampled), for k-space being undersampled using (x.1.) 1D variable density sampling taking 25% of the k-space, (x.2.) 2D variable density sampling taking 25% of the k-space, and (x.3.) 2D variable density sampling taking 10% of the k-space. SSIM values for (A.n.) magnitude and (B.n.) phase are presented.

Back to article page