PtyRANNOSAUR uses convolutional autoencoders trained on crystal structure databases to map 4D-STEM ptychography data to sub-0.5 Å phase images 10-100x faster than iterative methods while handling partial coherence, multiple scattering, and scan errors.
Ondry and Viraj Bodiwala and Peter Ercius and A
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.mtrl-sci 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
PtyRANNOSAUR: Ptychography with Robust Artificial Neural Networks Optimized for Sub-Angstrom Accuracy and Ultrafast Reconstruction
PtyRANNOSAUR uses convolutional autoencoders trained on crystal structure databases to map 4D-STEM ptychography data to sub-0.5 Å phase images 10-100x faster than iterative methods while handling partial coherence, multiple scattering, and scan errors.