We propose an adaptive algorithm for sparse image approximation, termed “adaptive thinning” (AT). The algorithm AT works with recursive removals of pixels from the target image, where the image is approximated by linear splines over anisotropic Delaunay triangulations. We discuss both computational and theoretical aspects of AT. The good performance of AT is finally supported by numerical examples and comparisons.

This presentation is part of Minisymposium “MS47 - Splines in Imaging (3 parts)”
organized by: Carolina Beccari (Dept. Mathematics, University of Bologna) , Virginie Uhlmann (EPFL, Lausanne) , Michael Unser (EPFL, Lausanne) .

Authors:

Armin Iske (Dept. Mathematics, University of Hamburg)

Keywords:

image compression, image representation, linear splines on triangulations, sparse approximation