Publication Details
Overview
 
 
, Luca Buzzatti, Thierry Scheerlinck, Jean-Pierre Baeyens, Michel De Maeseneer, , Erik Cattrysse, Gert Van Gompel, Jef Vandemeulebroucke,
 

Unpublished contribution to conference

Abstract 

Dynamic CT presents the potential to study joints while they undergo motion [1]. Kinematic parameters can also be extracted automatically/semi-automatically from these images using image processing techniques like registration. However, few studies have evaluated the accuracy of the motion parameters extracted via image registration. In this study, a custom-made circular polymethylmethacrylate (PMMA) phantom was designed to undergo controlled rotations. We implemented a two-step image registration approach to recover the induced rotations. Our registration approach was capable of recovering the true rotations with a mean absolute error of 0.52º [95% CI:0.36º-0.68º].

Reference