Abstract:Dual energy CT (DECT) expands applications of CT imaging in its capability to acquire two datasets, one at high and the other at low energy, and produce decomposed material images of the scanned objects. Bayesian theory applied for statistical DECT reconstruction has shown great potential for giving the accurate decomposed material fraction images directly from projection measurements. It provides a natural framework to include various kinds of prior information for improved image reconstruction with its optimal selected hyper parameter by a trial-error style. To eliminate the cumbersome style, in this work, we propose a parameter-free Bayesian reconstruction algorithm for DECT (PfBR-DE). In our approach, the physical meaning of the hyper parameter can be interpreted as the ratio of the data variance α and the prior tolerance σ by formulating the probability distribution functions of the data fidelity and prior expectation. With an alternative optimization scheme, the data variance, prior tolerance and decomposed material images can be jointly estimated. Experimental results with the abdomen phantom demonstrate the PfBR-DE method can obtain the comparable quantity decomposed material images with the conventional methods without freely adjustable hyper parameter.

Parameter-free Bayesian Reconstruction for Dual Energy Computed Tomograpy