Activity Number:
|
58
|
Type:
|
Invited
|
Date/Time:
|
Monday, August 12, 2002 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistical Computing*
|
Abstract - #300248 |
Title:
|
Structural Adaptation Methods in Imaging
|
Author(s):
|
Joerg Polzehl*+ and Vladimir Spokoiny
|
Affiliation(s):
|
Weierstrass Institute for Applied Analysis and Stochastics and Weierstrass Institute for Applied Analysis and Stochastics
|
Address:
|
Mohrenstr. 39, D-10117, Berlin, , D-10117, Germany
|
Keywords:
|
structural assumptions ; functional MRI ; adaptive smoothing ; image reconstruction ; signal detection
|
Abstract:
|
Images are often characterized by their discontinuities and homogeneous regions. Polzehl and Spokoiny (2000, JRSS B) propose a iterative nonlinear smoothing procedure that allows for adaptation with respect to the structural assumption of an local constant image. The method is based on locally constant smoothing with an adaptive choice of weights for every pair of data-points. Generalizations of the approach allow for local polynomial structures and general loss function. Applications include denoising of satellite (SAR) images as well as structuring of seismic geological data. The same idea has been successfully applied to construct a spatially adaptive procedure for signal detection in functional magnetic resonance imaging (Polzehl and Spokoiny (2001, JRSS C).
|