Iterative reconstruction techniques in low dose computed tomography applications

Abstract

The aim of this project will be to evaluate the efficacy of iterative reconstruction techniques in low dose computed tomography (CT) applications. Phantom studies will be performed to characterise noise and resolution properties of images acquired at different dose levels using traditional filtered-back-projection and iterative reconstruction techniques. Anthropomorphic phantoms will be used to quantitatively and visually assess images of clinically realistic structures formed using both the traditional and iterative reconstruction techniques across a range of doses. Novel analysis software will be created for the assessment of the visualization of these clinically realistic structures, including quantitative and observer-modelling techniques. Finally, findings will be verified using existing clinical scans and verified software that simulates low-dose acquisition by the addition of noise to existing raw CT acquisition data. Outcomes would include identification of the task-specific lower dose limits where iterative reconstruction algorithms can be viably used, lowering the radiation dose to patients undergoing clinical CT, and development of novel image assessment software for assessment of clinical image detail in CT.