Nonlinear PDEs in Image Restoration
by S.H. Lui, University of Manitoba

In the first half of this talk, we shall review the use of nonlinear
PDEs in denoising digital images.  Schemes of Perona-Malik,
Osher-Rudin, and mean curvature flow will be examined.

In the second half, some recent work on using Gaussian-curvature driven flow
by Lee and Keun will be discussed.  An embarassingly parallel domain
decomposition algorithm will be presented.  Numerical experiments will
compare this new scheme with other schemes.  Time permitting, some recent
work on deblurring will be presented.  This work is joint with Dmitry
Firsov.