Extended Data Fig. 8: SIM simulations with various types and magnitudes of motion.
From: Neural space–time model for dynamic multi-shot imaging
From left to right: a, rigid motion - translation, b, rigid motion - rotation, c, non-rigid global motion - shearing, and d, local deformable motion - swirl. The first four rows show the NSTM reconstructions from simulated images with increasing magnitude of motion between frames, and the last row shows the groundtruth scenes. The reconstruction of local deformable motion is more likely to fail when the motion magnitude increases. Full videos of the dynamic reconstructions can be found in Supplementary Video 7.
