Holotomocupy

Holotomography is a coherent imaging technique that provides three-dimensional reconstruction of a sample’s complex refractive index by integrating holography principles with tomographic methods. This approach is particularly suitable for micro- and nano-tomography instruments at the latest generation of synchrotron sources.

This software package presents a family of novel algorithms, encapsulated in an efficient implementation for X-ray holotomography reconstruction.

Features

• Based on Python, GPU acceleration with cuPy (GPU-accelerated numPy). Easy to install with pip, no C/C++ or NVCC compilers needed.

• Regular operators (tomographic projection, Fresnel propagator, scaling, shifts, etc.) and processing methods are implemented and can be reused.

• Jupyter notebooks give examples of full pipelines for synthetic/experimental data reconstruction.

• New operators/processing methods can be added by users. Implemented Python decorator @gpu_batch splits data into chunks if data do not fit into GPU memory.

• Pipeline GPU data processing with CUDA streams within cuPy allows significantly reduced time for some CPU-GPU memory transfers.

• Demonstrated for:

  1. Holotomography reconstruction with illumination retrieval.

  2. Holotomography reconstruction with coded apertures.

Contribute

• Documentation: https://github.com/tomography/holotomocupy/tree/master/doc

• Issue Tracker: https://github.com/tomography/holotomocupy/docs/issues

• Source Code: https://github.com/tomography/holotomocupy/

Content

• Installation
  • Update
• Usage
  • Synthetic data reconstruction
  • Experimental data reconstruction
  • Holotomography reconstruction with coded apertures
• API reference
  • holotomocupy.chunking
  • holotomocupy.holo
  • holotomocupy.magnification
  • holotomocupy.proc
  • holotomocupy.recon_methods
  • holotomocupy.shift
  • holotomocupy.tomo
  • holotomocupy.utils
• Credits
  • Citations
  • References