Joint direct motion estimation/kinetic images reconstruction from gated PET Data

Authors

Alexandre B. Bousse, University College London
Jieqing Jiao, University College London
Kris Thielemans, University College London
David Atkinson, University College London
Simon Arridge, University College London
Sebastien Ourselin, University College London
Brian F. Hutton, University of Wollongong, University College London

RIS ID

101587

Publication Details

Bousse, A., Jiao, J., Thielemans, K., Atkinson, D., Arridge, S., Ourselin, S. & Hutton, B. F. (2015). Joint direct motion estimation/kinetic images reconstruction from gated PET Data. In F. Gao, K. Shi & S. Li (Eds.), Computational Methods for Molecular Imaging (pp. 53-62). Switzerland: Springer International Publishing.

Abstract

In this paper we propose a unified framework for joint motion estimation/kinetic image reconstruction from gated dynamic PET data. The method is a generalisation of previous work to include gated data. The kinetic and motion parameters are estimated jointly by maximisation of the penalised likelihood. Kinetic parameters are estimated with an optimisation transfer approach, and the non-rigid motion is estimated with a quasi-Newton algorithm. Results on synthetic phantom data show that there is an advantage in jointly estimating motion and kinetics compared to pre-estimating the motion field for motion-compensated kinetic image reconstruction.