Preprint

An Investigation of Stochastic Variance Reduction Algorithms for 3D Penalised PET Image Reconstruction

Robert TwymanInstitute of Nuclear Medicine, UCL,Simon ArridgeInstitute of Nuclear Medicine, UCL, UKŽeljko KeretaMartinos Center for Biomedical Imaging, Harvard Medical School. S. Ahn is with GE Research, Niskayuna, NY 12309 USA. C. W. Stearns and F. Kotasidis are withBangti JinInstitute of Nuclear Medicine, UCL,Ludovica BrusaferriGE Healthcare, Waukesha, WI 53188 USA. ISangtae AhnInstitute of Nuclear Medicine, UCL, UKC.W. StearnsInstitute of Nuclear Medicine, UCL, UKBrian HuttonNuclear Medicine Department, University Hospital Zurich, University of ZurichIrene A. BurgerMartinos Center for Biomedical Imaging, Harvard Medical School. S. Ahn is with GE Research, Niskayuna, NY 12309 USA. C. W. Stearns and F. Kotasidis are withFotis A. KotasidisDepartment of Computer Science, UCL, UKKris ThielemansDepartment of Computer Science, UCL, UK

2022en

ABI

Abstract

Application of stochastic variance reduction algorithms to iterative PET reconstruction. We investigated the SAGA and SVRG algorithms for non-TOF PET image reconstruction. Both similated data and a patient data sets were used in the analysis. We found that the stochastic algorithms can improve convergence rate and eliminate behaviour, commonly know as limit cycle behaviour, from PET reconstruction within 5 epochs.

Topics

Medical Imaging Techniques and Applications Radiomics and Machine Learning in Medical Imaging Advanced MRI Techniques and Applications

Identifiers

DOI: 10.36227/techrxiv.19497680

Citations and references

Cited by 048 references

Metrics — AkademScholar