Title :
Parallel skeletonization by directional information
Author :
Frucci, Maria ; Marcelli, Angelo
Author_Institution :
Istituto di Cibernetica, CNR, Napoli, Italy
Abstract :
The paper presents a parallel algorithm to obtain the 8-connected labelled skeleton of a binary picture. It performs both the distance transformation and the erosion of the figure at the same time, and adopts directional information to guide such processes. The combination of directional and distance information leads to a reversible skeletonization which preserves both topological and geometrical properties of the figure, is largely rotation invariant, and exhibits a low computational cost
Keywords :
computational complexity; image reconstruction; image representation; iterative methods; parallel algorithms; 8-connected labelled skeleton; binary picture; computational cost; directional information; distance transformation; figure erosion; geometrical properties; parallel algorithm; reversible skeletonization; rotation invariant; topological properties; Computational efficiency; Euclidean distance; Iterative algorithms; Object recognition; Parallel algorithms; Shape; Skeleton;
Conference_Titel :
Image Processing, 1994. Proceedings. ICIP-94., IEEE International Conference
Conference_Location :
Austin, TX
Print_ISBN :
0-8186-6952-7
DOI :
10.1109/ICIP.1994.413401
