Data Description
1) Response_10.h5
a) Response matrix computed by GEANT4 for scintillator stack as described in the article, a series of monoenergetic photons where simulated and the energy deposited per photon in MeV recorded for each crystal
b) Stored in h5 format
2) hxLAS_inputdata.py
a) Python file with methods necessary to recreate the “bestguess” routine results in this article as an example for the full process. This file contains methods to load and interpolate the response, convert the response into single temperature responses necessary for the sparse/fine methods explained in the literature, and generates the input parameters used in comparison.
b) Requires libraries outlined below
o Pandas
o Numpy
o Matplotlib
o Scipy.Interpolate, Scipy.optimise, Scipy.Integrate
o CSV
o h5py
o lmfit
o datetime
o sys
3) “unguided” directory
a) Input Parameters
i) Csv of 250,000 input parameters stored as idx,n1,t1,n2,t2 for the unguided routine example (Figure 5)
b) Reconstructed parameters
i) Csv of 250,000 reconstructed parameters stored as idx,time,merit,Rn1,Rt1,Rn2,Rt2 for the unguided routine example (Figure 5,6)
4) “bestguess”, ”sparsefine”, and ”sparseopt” directories
a) Each contains:
i) Input Parameters
(1) Csv of 5,000 input parameters stored as idx,n1,t1,n2,t2 for the designated routine example (Figure 5,6,7)
ii) Reconstructed parameters for different values of noise
(1) Each CSV contains 5,000 reconstructed parameters stored as idx,time,merit,Rn1,Rt1,Rn2,Rt2. Noise values of 0, 0.1, 0.2, 0.5, 1, 1.5, 2, 3, 4, 5, 7.5, 10, 15, 20, 25 % were simulated. (Figure 5,6,7)
5) Reconstruction Equations - Supplementary Materials.nb
a) Mathematica notebook detailing the solution to reconstruction equations for N=1,N=2,N=3 as described in the text. Solutions are explicitly given for cases where X = M, and X=dM
b) To run this file use the software found at this URL for the
Wolfram Player for Notebooks: https://www.wolfram.com/player/