Welcome to

a PYTHon project dedicated to Iterative Reconstruction as it is used in the field of computed tomography.

Feel free to browse and/or clone the code yet keep in mind that this project currently can be considered a 'sandbox' of mine for several programming techniques such as

• GUI programming using PyQt and pyqtgraph
• GPGPU programming using PyOpenCl
• general pythonic coding (using Python 3 at some point...)
• testing using the unittest module
• experimenting with Cython or PyPy (maybe) and for several aspects of X-ray computed tomography such as
• definition of phantom data
• simulation of X-ray scans
• algebraic and statistical reconstruction methods
• forward/backprojection, system matrix My goal is to get self-taught in the techniques and principles listed above while keeping the overall project simple. F.i. I don't plan on implementing something fancier than a parallel beam geometry. But who knows.

As you can guess, this repo has some dependencies:

• pyqt4
• pyqtgraph
• numpy, scipy, matplotlib
• pyopencl

I browsed the web to find related work:

• CTSim: Appears to be abandoned. Clear and detailed documentation though. Comes with GUI.
• STIR: Actively developed C++ framework for iterative reconstruction of SPECT/PET data. Currently not funded anymore. Good documentation. Command line use only.
• GATE: Simulation package for PET/SPECT/CT. No reconstruction.

• DESIGN

• consistent backend data container for simulation
• cleanup QThread handling
• consistent handling of properties and pythonic methods
• introduce Forward/Backprojector methods
• handling of global settings like PyOpenCl usage

• COMPUTATIONAL METHODS

• ray tracing algorithm
• ML algorithm