Collaborative Computational Project in Synergistic Reconstruction for Biomedical Imaging

Synergistic Symposium 2019 X-rays

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

A. Perelli Spectral CT multi-material regularized reconstruction using randomized newton method An efficient and accurate randomized second-order algorithm for model-based image reconstruction is proposed and applied to spectral X-ray Computed Tomography with photon counting detectors. We consider the multi-material image model with a pixel-wise separable convex surrogate and each basis material is regularized by a data-driven prior. We propose to reduce the computational complexity using a partial randomized Hessian sketching, using the ridge leverage scores, only for the convex likelihood function and the regularization is designed using the denoising score matching framework which retains the complex prior structure. Finally, we show how to compute the gradient and the Hessian of the likelihood and regularizer together with simulated results.
Webster Stayman Model-based material decomposition for spectral CT Model-based iterative reconstruction has been widely adopted for tomography - first in nuclear imaging and more recently in x-ray CT. With the emergence of spectral CT where multiple energy channels are used to perform material decomposition, there is also growing work on the development of model-based methods to estimate material density maps directly from projection data (as opposed to staged estimation approaches involving reconstruction then decomposition, or vice versa). Such direct model-based methods enable novel acquisitions poorly suited to traditional processing approaches, but also introduce new complications in how to obtain the best estimates. This talk will introduce one variant of model-based material decomposition, illustrate the application for novel data acquisitions, and discuss analysis to predict and control image properties of the material density maps.
Guobao Wang PET-enabled spectral CT imaging using synergistic reconstruction Spectral CT acquires energy-dependent tissue attenuation information to allow quantitative multi-material decomposition. Its combination with PET has the potential to improve the capabilities of quantitative and multiparametric imaging. Standard methods for spectral CT imaging, however, commonly use two or more x-ray energies, either requiring a costly hardware change or significantly increasing radiation exposure if integrated with PET/CT. We propose a novel spectral CT imaging method that is enabled using time-of-flight PET/CT imaging without changing the hardware or increasing radiation dose. The PET-enabled spectral CT (PS-CT) uses the x-ray CT to provide low-energy (