eTomo

Electron tomography toolbox

eTomo is a python toolbox made for electron tomography

The basic forward and backward projection operations are based on the GPU-accelerated ASTRA-toolbox http://www.astra-toolbox.com/

This toolbox provides tools and reconstruction algorithms frequently used for building a electron tomographic pipeline, mainly for materials science.

Installations

Requirements/Dependency

• Numpy for data processing
• matplotlib for data visualization
• The astra-toolbox with GPU support. For installation of astra-toolbox, refer to http://www.astra-toolbox.com/
• Python environment: 64-bit 3.5 or 3.6
• Cuda 5.5 or hihger

Optional packages (might be used in the example codes)

• HypersPy: a library to process EDS/EELS spectroscopic images
• Operator Discretization Library (ODL): an operator-based package for optimization methods

Conda:

TODO

pip (for development)

Download the package, change to the source folder, type

pip install -e .

in the terminal

Functions:

Data I/O:

Support reading .mrc files. .mrc is a standard file format for tomographic measurement data from electron microscopes.

Preprocessing:

Basic functions are provided: Normalization: Alignment:

Data visualization

A slice viewer is provided basd on matplotlib. It is useful for viewing 3D volume / tilt-series of projection images Support viewing and ploting the reconstruction while the iterative reconstruction algorithm is running

Reconstruction algorithms

Default reconstruction algorithms implemented as follows:

• Analytical algorithms: FBP (simple and fast)
• Numerical algorithms: SIRT (smooth reconstructions)
• Statistical algorithms: EM (for data with strong Poisson noise)
• Advanced primal-dual algorithms: Chambolle-Pock It is also possible to implement your own reconstruction algorithm

Post-processing

TODO

Examples:

Examples

• TV-regularized STEM tomographic reconstruction
• Customized reconstruction algorithm
• HAADF-EDS bimodal tomographic reconstruction
• Nonlinearity correction for STEM tomography ...

Acknowledgment:

This software is a product of the computational imaging group at CWI, Amsterdam https://www.cwi.nl/research/groups/computational-imaging

Copyright: 2018, CWI, Amsterdam http://www.cwi.nl/