Multi-Scale Shape Description in Medical Imaging
Julia A. Schnabel, Arridge · 1997
. We have developed a hierarchical tool for shape description in medical imaging, based on the concepts of scale space and active contour models. The scale space representation allows to capture global shape outlines at a high level of scale or low image resolution, while finer scale shape details can be tracked for increasing levels of image resolution or decreasing scales, thus allowing to quantify shape changes across scales. The tracking process is implemented using active contour models, a technique which is formulated as active shape focusing. The tool has been applied to both artificial images and Magnetic Resonance(MR) images in clinical neurology, of which results will be presented and discussed. Keywords: Shape interpretation, active shape focusing, multi-scale shape stack, multi-scale active shape description, active contour models. 1 Introduction and Background Scale-space [10] techniques have recently attracted much interest in image processing: multi-scale segmentation a...