Innovative data weighting for iterative reconstruction in a helical CT security baggage scanner

Abstract

X-ray computed tomography (CT) currently has widespread application in air travel security systems for the purpose of baggage screening. This work presents an implementation of a fully 3D model-based iterative reconstruction (MBIR) algorithm mapped to a multi-slice helical CT security scanner. We introduce innovations in the data model that are designed to enhance image quality for typical scenes encountered in the security setting. In particular, we explore alternatives in the weighting of the measurements in order to more accurately reconstruct uniform regions and suppress metal artifacts. We compare images from the model-based approach to direct analytical reconstructions, indicating that the MBIR produces higher resolution and lower-noise reconstructions with suppressed metal streaking. The image improvements afforded by MBIR can provide for better operator experience and potentially enable enhanced performance of automatic threat detection (ATD).