Decomposing Digital 3D Shapes Using a Multiresolution Structure
In many applications, e. g. object recognition, decomposition of a shape is of great interest. We present a decomposition algorithm for 3D shape that is based on a multiresolution structure. The shape is hierarchically decomposed according to local thickness. A merging process is introduced for merging of small components to more significant parts. As a side effect of the algorithm, we also obtain a way of smoothing noisy shapes.
KeywordsMerging Process Local Thickness Distance Transform Merging Algorithm Pattern Recognition Letter
- 5.G. Borgefors and G. Sanniti di Baja. Parallel smoothing and decomposition of digital shapes using a multiresolution structure. In Proceedings 10th International Conference on Pattern Recognition, pages 745–748. IEEE Computer Society Press, 1990.Google Scholar
- 6.G. Borgefors, G. Sanniti di Baja, and S. Svensson. Multiresolution representation of shape in binary images ii: Volume images. In E. Ahronovitz and C. Fiorio, editors, Discrete Geometry for Computer Imagery (DGCI’97), pages 75–86. Springer Verlag, Berlin Heidelberg1997. Lecture Notes in Computer Science1347.CrossRefGoogle Scholar
- 11.K. Siddiqi, K.J. Tresness, and B.B. Kimia. On the anatomy of visual form. In C. Arcelli, L.P. Cordella, and G. Sanniti di Baja, editors, Aspects of Visual Form Processing, pages 507–521. World Scientific Publishing Co. Pte. Ltd., 1994.Google Scholar