Tiling-based Fresnel multislice propagation

This repo contains the codes for tiling-based Fresnel multislice propagation, which was demonstrated and benchmarked in the following publication(s):

Ali, S., Du, M., Adams, M., Smith, B., & Jacobsen, C. A comparison of distributed memory algorithms for x-ray wave propagation in inhomogeneous media. Opt Exp, (2020). (Submitted)

To reproduce the results shown in the paper, please switch to branch mpi_conv_bp, and run the following scrpts in directory cnn_propagator:

benchmark_pfft_zp.py is used for convergence test on the zone plate object (or any axially repeating objects).
benchmark_pfft.py is used for convergence test on the charcoal object. Please contact the authors for raw data.
s_scaling_pfft_zp.py is used for zone plate scaling test.
s_scaling_pfft_zp_gpu.py is used for GPU benchmarking, which requires PyTorch and Adorym (https://github.com/mdw771/adorym_dev).

Adorym (Automatic-Differentiation-based Object Retrieval using dYnamical Modeling) (THIS REPO IS OUTDATED)

Optimization-based 3D reconstruction algorithm for objects beyond the depth-of-focus using Tensorflow as an iterative engine

This code repo was used in the following publication(s):

Du, M., Nashed, Y. S. G., Kandel, S., Gürsoy, D. & Jacobsen, C. Three dimensions, two microscopes, one code: Automatic differentiation for x-ray nanotomography beyond the depth of focus limit. Sci Adv 6, eaay3700 (2020).

This is NOT the updated repo for Adorym. To retrieve the latest version of the software, visit https://github.com/mdw771/adorym_dev (developing repo) or https://github.com/mdw771/adorym (stable version).