Premium
SU‐E‐T‐75: Comparison of Gold Deconvolution and Backward Stripping for Correcting Low‐Energy Bremsstrahlung X‐Ray Spectra
Author(s) -
Moga J,
DeWerd L
Publication year - 2011
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.3612026
Subject(s) - deconvolution , bremsstrahlung , spectral line , detector , physics , computational physics , monte carlo method , energy (signal processing) , optics , mathematics , statistics , photon , quantum mechanics , astronomy
Purpose: To compare the iterative Gold deconvolution and backward stripping for correcting low‐energy bremsstrahlung x‐ray spectra. Methods: Moderately‐filtered x‐ray spectra (20kVp–250kVp) matched to the NIST M‐series beam qualities were measured with a low‐energy germanium detector. A full model of the spectrometry system was developed using the MCNP5 Monte Carlo code. The model was used to simulate measured spectra using sample bremsstrahlung input spectra and to calculate a detector response function (DRF). The iterative Gold deconvolution and backward stripping method were used with the MCNP5‐calculated DRF to correct the measured and simulated measured spectra. The effects of changing the size of the DRF and incorporating Gaussian energy broadening in the DRF were considered for both methods. The number of iterations was also varied for the deconvolution method. Results: Both methods were able to correctly determine the general shape of the bremsstrahlung spectra; however, differences between the two methods were observed at the high‐energy endpoints and in the low‐energy regions. For the deconvolution method, a continuum was observed at the high‐energy endpoint of the spectrum. The deconvolution was found to be sensitive to changes in the DRF size and the number of iterations, and the best results were achieved with no more than half the broadening modeled in the DRF. Backward stripping produced large statistical fluctuations in the extreme low‐energy region but a finite high‐energy endpoint. The stripping method was independent of DRF size and produced good results provided the amount of broadening was relatively small. Conclusions: Both the Gold deconvolution and backward stripping can be used to correct bremsstrahlung x‐ray spectra; however, more variability was observed when the number of iterations and DRF size were varied for the deconvolution. Backward stripping is considered a better choice for correcting bremsstrahlung x‐ray spectra.