Three-dimensional stylization of structures of interest from computed tomography images applied to radiotherapy planning.

Abstract

Techniques to detect structures of interest and model them as simple geometric shapes to enhance their visualization are presented. Three-dimensional visualization techniques can be used to enhance the detection of critical structures by enclosing them in a readily perceived bounding volume. This article describes stylization techniques that define three-dimensional shapes that conform to the real shape of the structures in the case of the spinal cord and the eyeballs. The term stylization can, thus, be defined as a combination of segmentation and the production of a tight bounding volume. The stylization techniques attempt to minimize the region that belongs to the solid produced and does not belong to the real structure, but they also attempt to ensure that the solid produced is a bounding volume to the real structure. The results produced by these techniques proved to be efficient and, thus, provide a means of easily visualizing the critical regions. The stylization technique can be seen as an adaptive volumetric contour. With potential use in clinical practice, it also provides a platform for further investigations of the geometric stylization concept.