BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Optimizing the optimizers - what is the right image and data model
 ? - Schnlieb\, C (University of Cambridge)
DTSTART:20140214T114500Z
DTEND:20140214T123000Z
UID:TALK50882@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:When assigned with the task of reconstructing an image from gi
 ven data the first challenge one faces is the derivation of a truthful ima
 ge and data model. Such a model can be determined by the a-priori knowledg
 e about the image\, the data and their relation to each other. The source 
 of this knowledge is either our understanding of the type of images we wan
 t to reconstruct and of the physics behind the acquisition of the data or 
 we can thrive to learn parametric models from the data itself. The common 
 question arises: how can we optimise our model choice?\n \nStarting from t
 he first modelling strategy this talk will lead us from the total variatio
 n as the most successful image regularisation model today to non-smooth se
 cond- and third-order regularisers\, with data models for Gaussian and Poi
 sson distributed data as well as impulse noise. Applications for image den
 oising\, inpainting and surface reconstruction are given. After a critical
  discussion of these different image and data models we will turn towards 
 the second modelling strategy and propose to combine it with the first one
  using a bilevel optimization method. In particular\, we will consider opt
 imal parameter derivation for total variation denoising with multiple nois
 e distributions and optimising total generalised variation regularisation 
 for its application in photography.\n \nJoint work with Luca Calatroni\, J
 an Lellmann\, Juan Carlos De Los Reyes and Tuomo Valkonen. \n\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR