dkazanc / larix Goto Github PK

Larix: is a set of data and image processing tools Larix is an open-source software written in C/CUDA languages with Python wrappers. The software consists of novel and already existing methods for various image processing tasks, e.g. filtering, inpainting, segmentation, morphological processing, etc. Most of the modules are implemented with OpenMP multithreading capability in C or on GPU to ensure faster data processing. Larix is being developed at Diamond Light Source synchrotron (UK) with the main purpose to process collected data effectively and efficiently.

Master	Anaconda binaries

• Mask evolving segmentation method (RegionGrow) with mask initialisation (2D/3D CPU)
• Morphological processing of the segmented image/volume (e.g. the result of RegionGrow) using the line segments to remove gaps and misclassified regions (2D/3D CPU)
• Auto cropping for tomographic projections or reconstructed images when the object is within the FOV (2D/3D CPU)
• Median filtration and median-based dezinger to remove broken pixels aka outliers (2D/3D CPU/GPU)
• Inpainting using linear/non-linear diffusion and non-local marching method (2D/3D CPU)
• A stripes detection approach

Install from the anaconda cloud: conda install -c dkazanc larix or build with:

export VERSION=`date +%Y.%m`
conda build recipe/ --numpy 1.15 --python 3.7
conda install -c file://${CONDA_PREFIX}/conda-bld/ larix

Additionally you can also run/modify conda_install.sh for automatic conda build/install

See run.sh for information how to build and install software.

• Savu Tomographic reconstruction pipeline at DLS to which Larix has been incorporated
• ASTRA-toolbox Versatile CPU/GPU toolbox for tomography
• ToMoBAR TOmographic iterative MOdel-BAsed Reconstruction software
• CCPi-RegularisationToolkit Adds regularisation to ToMoBAR
• TomoPhantom Tomographic projections and phantoms simulator