BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Randomized second-order algorithm for MAP estimation and fast MSE 
 estimator applied to Computed Tomography - Alessandro Perelli (DTU)
DTSTART:20191001T120000Z
DTEND:20191001T130000Z
UID:TALK131263@talks.cam.ac.uk
CONTACT:Carola-Bibiane Schoenlieb
DESCRIPTION:Two families of algorithms for MAP and MSE estimation respecti
 vely will be presented and applied to a monoenergetic X-ray Computed Tomog
 raphy (CT) acquisition model.\nSecond order methods for solving regularize
 d optimization problems with generalized linear models have been widely st
 udied but despite the superior convergence rate compared to first order me
 thods one weakness relies on the computational cumbersome for calculating 
 the Hessian matrix. Additionally\, in imaging applications where the input
  prior is difficult to model\, powerful regularization techniques are base
 d on data-driven models or denoisers.\nFor MAP estimation\,  an efficient 
 and accurate randomized second order method for model based CT reconstruct
 ion is proposed. The algorithm combines the idea of dimensionality reducti
 on of the Hessian of the likelihood cost function by sketching\, using rid
 ge leverage scores\, and an explicit regularizer term which can be impleme
 nted by a generic denoiser through the score matching formulation. We show
  how to compute the gradient and the Hessian of the likelihood and regular
 izer together with simulated results.\nFinally\, a fist order iterative me
 thod\, called approximate message passing\, will be presented for performi
 ng MSE estimation efficiently.
LOCATION:MR 14\, Centre for Mathematical Sciences
END:VEVENT
END:VCALENDAR