BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Deep Learning on geometric data - Davide Boscaini\, Università de
 lla Svizzera Italiana
DTSTART:20160204T111500Z
DTEND:20160204T121500Z
UID:TALK63768@talks.cam.ac.uk
CONTACT:Flora Tasse
DESCRIPTION:The past decade in computer vision research has witnessed the 
 re-emergence of "deep learning" and in particular\, convolutional neural n
 etwork techniques\, allowing to learn task-specific features from examples
  and achieving a breakthrough in performance in a wide range of applicatio
 ns. However\, in the geometry processing and computer graphics communities
 \, these methods are practically unknown. One of the reasons stems from th
 e facts that 3D shapes (typically modeled as Riemannian manifolds) are not
  shift-invariant spaces\, hence the very notion of convolution is rather e
 lusive. In this talk\, I will show some recent works from our group trying
  to bridge this gap. Specifically\, I will show the construction of intrin
 sic convolutional neural networks on meshes and point clouds\, with applic
 ations such as defining local descriptors\, finding dense correspondence b
 etween deformable shapes and shape retrieval. 
LOCATION:SS03 Meeting Room\, Computer Laboratory
END:VEVENT
END:VCALENDAR