Dual Estimation of Activity Maps and Kinetic Parameters for Dynamic PET Imaging

jingjia xu,huafeng liu,pengcheng shi,fei gao
DOI: https://doi.org/10.1007/978-3-319-18431-9_8
2015-01-01
Abstract:Due to rapid adoption of dynamic PET, the dual estimations of the activity maps and kinetic parameters have been attracting a lot of attention. In this paper, we propose a novel approach to solve this problem by using Dual-Kalman Filter (DKF) based on state space framework. One Kalman Filter is adopted to reconstruct the activity maps and the other is to estimate the kinetic parameters, where each filter uses the estimation results from the other one as initialization, then the two filters are solved iteratively until convergence. In addition, this approach combines the compartmental model guided activity map reconstruction and the state space based kinetic parameter estimation. The simulation experiments are presented by both utilizing DKF and other methods based on fitting the compartmental model. The final results show the more robust and accurate performance using proposed method.
What problem does this paper attempt to address?