Figure 2.

Unsupervised reconstruction using a coordinate network (CN) to create a 3D volume estimate, $\widehat{\mathbf{v}}$. This diagram depicts one iteration of training process, where network weights $\mathit{\theta }$ are updated by constraining the estimated projections $\widehat{\mathbf{p}}$ to match the given projections, $\mathbf{p}$. After the training process, $\widehat{\mathbf{v}}$ closely matches the ground-truth tomogram.