• home
• current students
• graduate
• events

Presenter: Travis Danniels
Supervisor: Prof. T. Aaron Gulliver

Date: Wed, November 20, 2013
Time: 11:00:00 - 00:00:00
Place: EOW 430

ABSTRACT

ABSTRACT:

Deblurring is the problem of recovering a sharp image from a blurry image. A deblurring problem may be categorized as non-blind, in which the blur kernel (point-spread function) is known, or blind, in which it is not. In this talk, an overview of deblurring is presented, and algorithms for blind and non-blind deblurring are described. The non-blind algorithm is based on a convex program consisting of a data fitting term and a sparsity-promoting regularization term. The data fitting term is the l2 norm of the residual between the blurred image and the latent image convolved with a known blur kernel. The regularization term is the l1 norm of the latent image under a wavelet frame (framelet) decomposition. This convex program is solved with the first-order primal-dual algorithm recently proposed by Chambolle and Pock. The blind deblurring algorithm is based on recent work by Cai, Ji, Liu, and Shen. It works by embedding the proposed non-blind algorithm in an alternating minimization scheme and imposing additional constraints in order to deal with the challenging non-convex nature of the blind deblurring problem. Numerical experiments on artificially and naturally blurred images are presented.