Department of Mathematics Colloquium


Abstract 

I will first briefly introduce multivariate
box spline functions which are straightforward generalization of
the wellknown Bsplines. Then I shall explain a constructive method to
find tight wavelet frames based on box splines. Among many of
these frames, I will present one of frames which is constructed
using an 8 directional box spline. The second part of my talk is to
show you some applications of this frame for image edge/segmentation
analysis. 1) accurately detection of image edges and comparison with
other methods will be given to demonstrate the advantage of
this frames. 2) application of this frame for Dirac edges and hidden
edges and 3) application for segmentation of images.
