ODT with Lippmann-Schwinger model

Description

This repository provides codes for 2D and 3D optical diffraction tomography (ODT) with iterative Lippmann-Schwinger model. More details can be found in the following paper:

[1] Efficient inversion of multiple-scattering model for optical diffraction tomography.,
Optics Express, 25, 21786–21800 (2017),
E. Soubies, T-a. Pham, and M. Unser.

[2] Three-Dimensional Optical Diffraction Tomography with Lippmann-Schwinger Model.,
IEEE Transactions on Computational Imaging (2020),
T-a. Pham, E. Soubies, A. Ayoub, J. Lim, D. Psaltis, and M. Unser.

[3] A method for assessing the fidelity of optical diffraction tomography reconstruction methods using structured illumination.,
Optics Communications, 454, 124486 (2020),
A. B. Ayoub, T-a. Pham, J. Lim, M. Unser, and D. Psaltis.

Requirements

The code requires the GlobalBioIm library v1.1.2 (or more recent releases)
https://biomedical-imaging-group.github.io/GlobalBioIm/

Repository content

The repository is organized as follows.

• Folder 2D: two-dimensional ODT [1]

  • In folder code, the script main_fig7.m reproduces Figure 7 of [1].
  • In folder code, the script ExampleFwdModel.m provides an example on how to use the iterative Lippmann-Schwinger forward model,
  • The folder mat_files contains data.

• Folder 3D: three-dimensional ODT [2]

  • The script DiscretizationGreen.m reproduces Figure 3 of [2],
  • The script ExampleFwdModel.m provides an example on how to use the iterative Lippmann-Schwinger forward model,
  • The script Compute_uin.m reproduces Figure 5 of [2],
  • The script ODT_Reconstruction.m contains the whole reconstruction pipeline.
  • The script Assessment_method.m implements the method of assessment of [3].

  Notes

  • The data for ODT_Reconstruction.m can be downloaded here
  • ODT_Reconstruction takes ~6h on CPU (Intel(R) Core(TM) i7-8650U CPU @ 1.90GHz).
  • ODT_Reconstruction takes ~1h30 on GPU (Nvidia GeForce GTX Titan X).