On a (W)ENO-Type Multiscale Representation Based on Quincunx Refinement: Application to Image Compression

Basarab Matei (1) , Sylvain Meignen (2) , Anastasia Zakharova (2)

1 LAGA - Laboratoire Analyse, Géométrie et Applications
2 MGMI - Modélisation Géométrique & Multirésolution pour l'Image

In this paper, we study nonlinear multiscale representations on ℝ2 which are interpolatory and based on non-diagonal dilation matrices, such as the quincunx matrix. A compression procedure is then introduced for that kind of representations while numerical experiments conclude the paper.

Mots clés: Nonlinear Multiscale Representation; Image Compression

Domaines: Analyse numérique [math.NA]

Basarab Matei, Sylvain Meignen, Anastasia Zakharova. On a (W)ENO-Type Multiscale Representation Based on Quincunx Refinement: Application to Image Compression. 7th International Conference on Curves and Surfaces, Jun 2010, Avignon, France. pp.473-487, ⟨10.1007/978-3-642-27413-8_30⟩. ⟨hal-00872506⟩