Comparison of different follow-up lung registration methods with and without segmentation

摘要: In modern multi slice CT scanners the increasing amount of data also increases demand on image processing methods that assist diagnosis. For detection and classification lung nodules in a follow up study it is very helpful to have slices previous scan aligned with current scan. This typical problem registration, for which different types solutions exist. We investigated accuracy computation times rigid body, an affine, spline based elastic registration approach complete set, compared results method where was preceded by segmentation volume. The quality determined ground truth previously nodule locations measuring average distance corresponding nodules. It found affine slightly better than body both are much faster turn showed best quality. A good compromise segmented volume, total not slower without segmentation, but shows alignment higher robustness.