Python implementation of the NN-FBP Algorithm, published in [1].
To install PyNN-FBP, simply run:
python setup.py install
To use PyNN-FBP, you need installed:
- Numpy and scipy
- PyTables
- (for ASTRAProjector) The ASTRA toolbox (https://code.google.com/p/astra-toolbox/), with Python interface (https://github.com/dmpelt/pyastratoolbox)
- (for SimplePyCUDAProjector) PyCUDA and pyfft.
Examples can be found in the examples directory. After installation, the examples can be run to show how the package is used.
Running PaperExample.py from the examples directory should show comparable results to the threeshape experiment (Fig. 9a and Table 1) of [1].
[1] Pelt, D., & Batenburg, K. (2013). Fast tomographic reconstruction from limited data using artificial neural networks. Image Processing, IEEE Transactions on, 22(12), pp.5238-5251.