Friday, October 13, at 327 Yost
Refreshments: 3:30 - 4:00 p.m, Talk: 4:00
- 5:00 p.m.
Blind deconvolution seeks to deblur an image without knowing the
cause of the blur. Iterative methods are commonly applied to that
problem, but the iterative process is slow, uncertain, and often
ill-behaved. This talk considers a significant but limited class
of blurs that can be expressed as convolutions of 2-D symmetric
Levy `stable' probability density functions. This class includes
and generalizes Gaussian and Lorentzian distributions. For such
blurs, methods are developed that can detect the point spread
function from 1-D Fourier analysis of the blurred image. A separate
image deblurring technique uses this detected point spread function
to deblur the image. Each of these two steps uses direct non-iterative
methods, and requires interactive adjustment of parameters. As a
result, blind deblurring of 512X512 images can be accomplished in
minutes of CPU time on current desktop workstations. Numerous blind
experiments on synthetic data show that for a given blurred image,
several distinct point spread functions may be detected that lead to
useful, yet visually distinct reconstructions. Application to real
blurred images will also be demonstrated.