Return to Colloquia & Seminar listing
Anisotropic Multi-Scale Estimation via Directional LPA-ICI Technique
Applied Math| Speaker: | Karen Egiazarian, Institute of Signal Processing of Tampere University, Finland |
| Location: | 2112 MSB |
| Start time: | Thu, Mar 30 2006, 4:10PM |
Description
The directional local polynomial approximation (LPA) is a technique applied
for a linear and nonlinear filter design using a polynomial fit in a sliding
window. The window size of this fit is one of the key-parameters interpreted
as a scale of adaptation and estimation.
It is a novel powerful tool for many signal/image/video processing problems:
analysis, denoising, deblurring, segmentation, differentiation, edge
detection, etc. In particular, in many imaging systems recorded observations
have the physical meaning of numbers of detected photons. The photons are
counted at different spatial locations and form an image of an object. This
sort of scenario is typical for so-called photon-limited imaging including
digital photo/video, positron emission tomography, astronomy, microscopy,
etc. It will be shown that the new approach demonstrates the state-of-art
performance and in many occasions visually and quantitatively outperform
some of the best existing methods.
Additional material is on the web page: www.cs.tut.fi/~lasip
Note special time for the Applied Math Seminar