Prof. Maks Ovsjanikov (École Polytechnique)
This talk discuss a representation of maps between pairs of 3D shapes (represented as triangle meshes) that generalizes the standard notion of a map to include correspondences that are not necessarily point-to-point. This representation is compact, and yet allows for efficient inference (shape matching) and enables a number of applications, including algebraic map manipulation such as computing map sums and differences.
The key aspect of this representation is that many constraints on a map, including landmark correspondences, part preservation and operator commutativity become linear. This means, in particular, that shape matching can be phrased as a simple linear system of equations.
I will describe the main properties of this representation and give a few examples of applications that include improving existing correspondence, segmentation transfer without establishing point-to-point matches and efficient map compression and visualization.
Prof. Ovsjanikov earned his MSc (2007) and PhD (2011) degrees in computational mathematics from Stanford University. He is now an Associate Professor at the computer science department in Ecole Polytechnique, Paris with a CNRS chair d'excellence. His PhD work on shape analysis was highly influential in the geometry processing field, received various best paper awards, as well as an Excellence in Research award given to a single Stanford graduating student. Prof. Ovsjanikov additionally interned at Adobe Creative Labs, eBay Research and Google, where he also spent a year in the Image Search team.