Title :
Fast skeletonization of spatially encoded objects
Author :
Romero, Francisco ; Ros, Lluis ; Thomas, Federico
Author_Institution :
CSIC, Barcelona, Spain
Abstract :
Some thinning algorithms for 3D objects, or generalizations of existing ones for 2D, have been proposed in recent years. The paper presents a simple and very fast algorithm compared to most of them, and still it has theoretically favorable properties. It provides a connected surface skeleton that allows shapes to be reconstructed with bounded error. In addition, it is also very attractive because it allows discrete skeletons to be obtained directly from volumes in many representations without converting them to a voxel-based representation. Our algorithm is a generalization of the one presented by Cardoner et al. (1997) for 2D objects. It is based on the application of directional erosions, while retaining those voxels that introduce disconnection
Keywords :
image reconstruction; image representation; image thinning; mathematical morphology; image reconstruction; image representation; mathematical morphology; skeletonization; spatially encoded objects; thinning; Animation; Biomedical imaging; Discrete transforms; Image reconstruction; Iterative algorithms; Robustness; Service robots; Shape control; Skeleton; Surface reconstruction;
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
Print_ISBN :
0-7695-0750-6
DOI :
10.1109/ICPR.2000.903595