A Hybrid Thinning Algorithm

His research interest includes medical image registration, fusion and 3D parallel thinning algorithms. Attila Kuba is the head of the Department of Applied Informatics, JJ ozsef Attila University , Szeged. His research interests are image processing, its applications in medicine, image reconstruction and discrete tomography. Thinning is a frequently used method for extracting skeletons in discrete spaces. This paper presents an eecient parallel algorithm for thinning elongated 3D binary objects (e.g., bony structures, vessel trees, or airway trees). The proposed algorithm directly extracts medial lines as shape features from 3D binary objects by applying a brand{new class of thinning strategy called hybrid method. Our topology preserving algorithm makes easy implementation possible and gives satisfactory results for synthetic data tests and for MR angiography brain studies. Introduction Skeletonization provides shape features that are extracted from binary image data. It is a common preprocessing operation in raster{to{vector conversion or in pattern recognition. Its goal is to reduce the volume of elongated objects to their skeletons. In the 3D Euclidean space, the skeleton of an object is the locus of the centers of all the maximal inscribed spheres (Blum, 1967). In discrete spaces, the thinning procedure is a frequently used method for generating an approximation to the continuous skeleton in a topology preserving way (Kong & Rosenfeld, 1989). Thinning is similar to peeling an onion. One step of the iterative method removes the outmost layer of an object and the entire process is repeated until only the \skeleton" is left. There are two major methods for shape representation. The rst method describes the boundary that surrounds an object. The second one gives a representation of the region that is occupied by the object to be analyzed. Boundary{based techniques are widely used but there are some deeciencies which limit their usefulness in practical applications 1 A Hybrid Thinning Algorithm. .. 2 especially in 3D (Sz ekely, 1996): methods of diierential geometry are rather sensitive to noisee occlusion may seriously disturb boundary{based descriptorss they are not appropriate to catch global shape features and to make them explicitt they can rather hardly reveal the hierarchical organization of the shape. The concept of skeletonization should beable to help exactly at the points listed above. The local object symmetries represented by t h e s k eleton certainly cannot replace boundary{ based shape descriptors, but complement a n d support them. The importance of the skeleton as a …