Fig. 3
From: A review on deep learning MRI reconstruction without fully sampled k-space
Image reconstruction with self-partition undersampled k-space data. Acquired undersampled k-space data \(\Omega\) will be divided into two subsets satisfying \(\Omega = \Lambda_{j} \cup \Theta_{j}\) before training network, where \(j = 1,\ldots, {\text{K}}\) denoting the number of partitions for each scan, \(\Lambda\) and \(\Theta\) is used as input for training and to calculate the loss function separately. The network is unrolled based on the VSQP algorithm. This figure is reproduced following Fig. 1 in Ref. [65]