Algorithms for Shape Matching and Approximation

The aim of this project is the development and partial implementation of algorithms for similarity determination and approximation of geometric objects. To achieve this, methods of computational geometry are to be applied in order to recognize and approximate patterns and shapes. Earlier works of the work group concerning this topic shall be generalized to higher dimensions and more general transformations for the matching of shapes, e.g. arbitrary affine mappings. In particular, data structures that allow to determine the most similar one out of a fixed set of shapes shall be developed. The setting into practice of the complex data structures and methods that most of the algorithms contain, as well as the application of approximation-approaches like for example reference-point-methods, shall also be examined.

More information