Legacy display Course

Spaces of Shapes as Riemannian manifolds - from modeling to efficient computation

Selected Topics in Scientific Computing (V5E4)

Lecturer: Prof. Dr. Martin Rumpf

The lecture course will discuss the space of images and the space of curves or surfaces from the perspective of Riemannian geometry. Examples are the flow of diffeomorphism and the optimal transportation approach in imaging as well as the space of viscous rods in geometric modeling. The course will discuss the different underlying metric structures and will study the existence of shortest connecting paths. Furthermore, a variational time discretization will be introduced and the convergence to the continuous model will be investigated. Finally, appropriate space discretizations and effective numerical algorithms will be presented.

Further information can be found here.