Statistical image reconstruction for polyenergetic X-ray computed tomography

Abstract

This paper describes a statistical image reconstruction method for X-ray computed tomography (CT) that is based on a physical model that accounts for the polyenergetic X-ray source spectrum and the measurement nonlinearities caused by energy-dependent attenuation. We assume that the object consists of a given number of nonoverlapping materials, such as soft tissue and bone. The attenuation coefficient of each voxel is the product of its unknown density and a known energy-dependent mass attenuation coefficient. We formulate a penalized-likelihood function for this polyenergetic model and develop an ordered-subsets iterative algorithm for estimating the unknown densities in each voxel. The algorithm monotonically decreases the cost function at each iteration when one subset is used. Applying this method to simulated X-ray CT measurements of objects containing both bone and soft tissue yields images with significantly reduced beam hardening artifacts.

Sparsity-based Sinogram Denoising for low-dose Computed Tomography

“…In the common statistical model of the CT scan, the measurements y are rendered as instances of the random variables Y ∼ P oisson(λ ), where λ = I 0 e −ḡ . Using a second-order approximation of the log-likelihood expression log(P(Y = y |f )) for this model, one deduces that the variance of the deviation between the measured sinogram valueĝ = −log(y /I 0 ), and the ideal valueḡ equals 1/λ [6,7]. In practice, this value is well approximated by 1/y .…”

High-Resolution CT for Small-Animal Imaging Research

“…Statistical image reconstruction is another class of iterative reconstruction algorithms, first not only introduced for transmission imaging in nuclear medicine but also applied for CT reconstruction (Elbakri and Fessler, 2002). In essence, these algorithms treat the reconstruction as a statistical estimation problem and have the advantage that they can take into account the Poisson model noise in projection data.…”

Adaptive filtered-back-projection for computed tomography

“…Desired properties of a reconstructed image can be introduced to PL function as a prior, thus improving the quality of reconstruction. We refer to the papers [5,2] for comprehensive description of such algorithms.…”