Simultaneous Reconstruction of Tissue Attenuation and Radioactivity Maps in SPECT

Abstract

Abstract. The importance of accurate attenuation correction in single photon emission computed tomography (SPECT) has been widely recognized. In this paper, we propose a novel scheme of simultaneous reconstruction of the tissue attenuation map and the radioactivity distribution from SPECT emission sinograms, which is obviously beneficial when the transmission data is missing for cost or efficiency reasons. Our strategy combines the SPECT image formation and data measurement models, whereas the attenuation parameters are treated as random variables with known prior statistics. After converting the models to state space representation, the extended Kalman filtering procedures are adopted to linearize the equations and to provide the joint estimates in an approximate optimal sense. Experiments have been performed on synthetic data and real scanning data to illustrate abilities and benefits of the method.

Nonlinear Dual Reconstruction of SPECT Activity and Attenuation Images

“…For these two sets of synthetic projection data, the radioactivity maps are reconstructed using four reconstruction methods: FBP [28] , EM-ML [29] , EKF [24] , and dual UKF (The attenuation maps are also reconstructed for the EKF and UKF methods). For the UKF framework, the state (activity map) is estimated by the UKF state-filter, where, for the first iteration, the attenuation parameters come from the initial guess.…”

“…Some of the most notable works include the use of differential attenuation method [17] , gradient ascent [18] , Tikhonov regularization [19] , and expectation maximization (EM) [20] – [23] . With the SPECT imaging and measurement processes in state space representation, a recent work has adopted the extended Kalman filtering (EKF) procedures to linearize the augmented state representation to provide the joint estimates in the minimum-mean-square-error sense [24] . Such EKF based framework, however, has several potential drawbacks.…”