Shape Analysis with Geometric Primitives
In this chapter, a unifying framework is presented for analyzing shapes using geometric primitives. It requires both a model of shapes and a model of geometric primitives. We deliberately choose to explore the most general form of shapes through the notion of connected components in a binary image. According to this choice, we introduce geometric primitives for straightness and circularity which are adapted to thick elements. Starting from a model (the tangential cover) which emerged in the digital geometry community we show how to use Constrained Delaunay Triangulation to represent all shapes as a connected path well adapted to the recognition of geometric primitives. We moreover describe how to map our framework into the class of circular arc graphs. Using this mapping we present a multi-primitives analysis which is suitable for self-organizing a shape with respect to prescribed geometric primitives. Further work and open problems conclude the chapter.
KeywordsPoint Index Steiner Point Open Node Geometric Primitive Constrain Delaunay Triangulation
- 2.Buzer, L.: Digital line recognition, convex hull, thickness, a unified and logarithmic technique. In: Proceedings of the 11th IWCIA, Berlin, Germany. Lecture Notes in Computer Science, vol. 4040, pp. 189–198 (2006) Google Scholar
- 7.Faure, A., Feschet, F.: Robust decomposition of thick digital shapes. In: Proceedings of the 12th IWCIA, Buffalo, USA. Lecture Notes in Computer Science, vol. 4958, pp. 148–159 (2008) Google Scholar
- 20.Prasad, L., Rao, R.: A geometric transform for shape feature extraction. In: Vision Geometry 2000. Proceedings of SPIE, vol. 4117, pp. 222–233 (2000) Google Scholar
- 21.Reveilles, J.-P.: Geometrie discrete, calcul en nombres entiers et algorithmique. These d’Etat (1991) Google Scholar
- 24.Vialard, A.: Geometrical parameters extraction from discrete paths. In: 6th International Workshop DGCI. Lecture Notes in Computer Science, vol. 1176, pp. 24–35. Springer, Berlin (1996) Google Scholar