TY - GEN
T1 - Improving a Stochastic Algorithm for Regularized PET Image Reconstruction
AU - Delplancke, Claire
AU - Gurnell, Mark
AU - Latz, Jonas
AU - Markiewicz, Pawel J.
AU - Schönlieb, Carola Bibiane
AU - Ehrhardt, Matthias J.
N1 - Funding Information:
Manuscript received December 19, 2020. This work was funded by the UK EPSRC grant PET++: Improving Localisation, Diagnosis and Quantification in Clinical and Medical PET Imaging with Randomised Optimisation EP/S026045/1.
Publisher Copyright:
© 2020 IEEE
PY - 2021/8/12
Y1 - 2021/8/12
N2 - Positron Emission Tomography (PET) image reconstruction presents challenges related to the large scale of data to be processed, which affects reconstruction speed, and the need to include regularizers to improve image quality. Among the methods proposed to overcome these challenges, the recently introduced Stochastic Primal Dual Hybrid Gradient (SPDHG) algorithm combines the ability to deal with regularizers like Total Variation and to process large datasets by random subsampling. We present two contributions regarding the step-sizes of SPDHG: i) larger step-sizes facilitated by a new formula, and ii) a numerical method to calibrate, in the context of PET reconstruction, the tradeoff between primal and dual progression, which is common to all primal-dual algorithms. We validate improvements in speed reconstruction on real PET data from the Siemens Biograph mMR.
AB - Positron Emission Tomography (PET) image reconstruction presents challenges related to the large scale of data to be processed, which affects reconstruction speed, and the need to include regularizers to improve image quality. Among the methods proposed to overcome these challenges, the recently introduced Stochastic Primal Dual Hybrid Gradient (SPDHG) algorithm combines the ability to deal with regularizers like Total Variation and to process large datasets by random subsampling. We present two contributions regarding the step-sizes of SPDHG: i) larger step-sizes facilitated by a new formula, and ii) a numerical method to calibrate, in the context of PET reconstruction, the tradeoff between primal and dual progression, which is common to all primal-dual algorithms. We validate improvements in speed reconstruction on real PET data from the Siemens Biograph mMR.
UR - http://www.scopus.com/inward/record.url?scp=85124704975&partnerID=8YFLogxK
U2 - 10.1109/NSS/MIC42677.2020.9508013
DO - 10.1109/NSS/MIC42677.2020.9508013
M3 - Chapter in a published conference proceeding
AN - SCOPUS:85124704975
T3 - 2020 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2020
BT - 2020 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2020
PB - IEEE
T2 - 2020 IEEE Nuclear Science Symposium and Medical Imaging Conference, NSS/MIC 2020
Y2 - 31 October 2020 through 7 November 2020
ER -