The Taub Faculty of Computer Science Events and Talks

The lecture deals with invariance in image processing, with emphasis on invariance of shapes and curves. There are two main topics. First, we consider scale invariance, review existing scale invariant measures and discuss the different ways in which an algorithm can be scale invariant. We then point out some dangers inherent in fully scale invariant algorithms. Second, we present an equi-affine invariant shape matching mathod, which is also invariant to resampling, robust to pixelisation noise and can perform under occlusion. The algorithm is a unification of ideas from semi-differential invariants, shape contexts, and shape distributions.