About
GENFIRE, for GENeralized Fourier Iterative REconstruction, is a software package implementing the algorithm of the same name for 3D image reconstruction from arbitrarily-oriented projection images. GENFIRE exists as both a Python implementation with a GUI programmed with PyQt5 and as a bundle of MATLAB code.
File formats / conventions
The Euler angle convention used by GENFIRE corresponds to that established by Heymann, Chagoyen and Belnap (2005). Each Euler angle triplet is expressed as (phi, theta, psi) where:
- phi is a rotation about the z axis
- theta is a rotation about the y' axis
- psi is a rotation about the z'' axis
This Euler angle data should be provided as a 2D array of dimension num_projections x 3 with one row per projection. This may be either a .npy, .mat, or .txt file delimited in a way that it can be read by default with numpy.loadtxt (space-delimited works).
The projection data should be a 3D volume of size N x N x num_projections containing floating-point precision data. N should be an even number, and currently must be the same for both the X and Y dimensions of the projections.
The location of the rotational center is defined to be at pixel N/2 with 0-based indexing as in Python. In MATLAB, which uses 1-based indexing, this pixel corresponds to (N/2 + 1).
Running reconstructions with Python
GENFIRE reconstructions may be run from within the genfire Python package through use of the GenfireReconstructor class. Simply instantiate a GenfireReconstructor with the desired reconstruction parameters and then invoke the reconstruct method. The reconstruction volume and associated error curves may then be saved using the helper function gf.fileio.saveResults. A simple example follows. The full list of possible options and their associated default values may be viewed with help gf.reconstruct.GenfireReconstructor.
import genfire as gf
pars = dict(
projectionFilename = "./data/projections.mat",
angleFilename = "./data/angles.mat",
supportFilename = "./data/support.mat",
resultsFilename = "recon.mrc",
numIterations=50,
)
GF = gf.reconstruct.GenfireReconstructor(**pars)
results = GF.reconstruct()
gf.fileio.saveResults(results, GF.params.resultsFilename)
Which, when run from within Python, might produce the following output
______ ________ ____ _____ ________ _____ _______ ________
.' ___ ||_ __ ||_ \|_ _||_ __ ||_ _||_ __ \ |_ __ |
/ .' \_| | |_ \_| | \ | | | |_ \_| | | | |__) | | |_ \_|
| | ____ | _| _ | |\ \| | | _| | | | __ / | _| _
\ `.___] |_| |__/ | _| |_\ |_ _| |_ _| |_ _| | \ \_ _| |__/ |
`._____.'|________||_____|\____||_____| |_____||____| |___||________|
projectionFilename = ./data/projections.mat
angleFilename = ./data/angles.mat
supportFilename = ./data/support.mat
resultsFilename = recon.mrc
resolutionExtensionSuppressionState = 1
numIterations = 50
oversamplingRatio = 3
interpolationCutoffDistance = 0.5
useDefaultSupport = True
calculateRfree = True
initialObjectFilename = None
constraint_positivity = True
constraint_support = True
enforceResolutionCircle = True
permitMultipleGridding = True
Reading projections from MATLAB file.
Assembling Fourier grid.
Fourier grid assembled in 1.3 seconds
Reconstruction started
Iteration number: 1 error = 1.00000
Iteration number: 2 error = 0.91111
Iteration number: 3 error = 0.89307
Iteration number: 4 error = 0.88159
Iteration number: 5 error = 0.56852
Iteration number: 6 error = 0.52332
Iteration number: 7 error = 0.34005
Iteration number: 8 error = 0.28827
Iteration number: 9 error = 0.26624
Iteration number: 10 error = 0.19585
Iteration number: 11 error = 0.16516
Iteration number: 12 error = 0.14891
Iteration number: 13 error = 0.11899
Iteration number: 14 error = 0.10121
Iteration number: 15 error = 0.09058
Iteration number: 16 error = 0.07963
Iteration number: 17 error = 0.07196
Iteration number: 18 error = 0.06201
Iteration number: 19 error = 0.05646
Iteration number: 20 error = 0.05293
Iteration number: 21 error = 0.05006
Iteration number: 22 error = 0.04824
Iteration number: 23 error = 0.04700
Iteration number: 24 error = 0.04615
Iteration number: 25 error = 0.04555
Iteration number: 26 error = 0.04512
Iteration number: 27 error = 0.04482
Iteration number: 28 error = 0.04460
Iteration number: 29 error = 0.04448
Iteration number: 30 error = 0.04441
Iteration number: 31 error = 0.04435
Iteration number: 32 error = 0.04440
Iteration number: 33 error = 0.04444
Iteration number: 34 error = 0.04450
Iteration number: 35 error = 0.04461
Iteration number: 36 error = 0.04473
Iteration number: 37 error = 0.04490
Iteration number: 38 error = 0.04507
Iteration number: 39 error = 0.04523
Iteration number: 40 error = 0.04543
Iteration number: 41 error = 0.04564
Iteration number: 42 error = 0.04584
Iteration number: 43 error = 0.04608
Iteration number: 44 error = 0.04631
Iteration number: 45 error = 0.04654
Iteration number: 46 error = 0.04680
Iteration number: 47 error = 0.04705
Iteration number: 48 error = 0.04731
Iteration number: 49 error = 0.04759
Iteration number: 50 error = 0.04787
Reconstruction finished in 16.4 seconds