From motion capture of interacting hands to video based rendering.

December 4, 2012
Luca Ballan | Institute of Visual Computing in ETH Zurich

The talk will be structured in two parts.
I will first talk about my research on marker-less motion capture, in particular presenting my recent ECCV work on hands tracking. Capturing the motion of hands interacting with each other and with an object is a challenging task due to the large number of degrees of freedom, self-occlusions, and similarity between the fingers. I will show how we addressed this problem by proposing a generative approach modeling all the elements of the scene as a unique articulated deformable object, and exploiting some discriminatively learnt features.
In the second part, I will talk about my research on interactive free-viewpoint video. I will show how it is possible to navigate a collection of videos representing an event casually captured by people in the audience. Instead of recovering a perfect 3D representation of the recorded scene, which in general is very challenging, I will propose several simplifying assumptions allowing video based rendering in such a scenario.
I will also show a live real-time demo of the proposed interactive navigation tool on different video collections.
URLs:
http://cvg.ethz.ch/research/ih-mocap/
http://cvg.ethz.ch/research/unstructured-vbr/
http://www.inf.ethz.ch/personal/lballan/

Speaker Details

Luca Ballan is a post doc at the Institute of Visual Computing in ETH Zurich. In 2005 he finished his computer engineering studies at the University of Padova, Italy, and in 2008 he received his PhD degree from the same university with a dissertation on marker-less motion capture. Luca’s research interests include passive 3D reconstruction of static and dynamic scenes, in particular estimating the shape and the pose of articulated objects from videos, and video based rendering.