The Taub Faculty of Computer Science Events and Talks

Comparison of deformable shapes is an important problem in computer vision, arising, for example, in the face recognition application. In this talk, we will present an axiomatic construction of similarity criteria for deformation-invariant shape comparison, based on intrinsic geometric properties of the shapes. Such criteria can be efficiently computed using a generalization of multidimensional scaling (GMDS). We will also address the setting of partially similar deformable shapes, i.e., shapes having large similar parts. We will present a generic approach to quantify partial similarity based on the notion of Pareto optimality. We will exemplify our approach in several applications, including the problems of recognition of non-rigid shapes like silhouettes and surfaces. (This seminar will be preceded by a Ph.D. seminar by Alexander Bronstein at 14:30).