Combining Local Regularity Estimation and Total Variation Optimization for Scale-Free Texture SegmentationMS78

Texture segmentation still constitutes an ongoing challenge, especially when processing large-size images. In this contribution we focus on (i) extracting simultaneously characteristics such as local regularity and local variance, integrating a scale-free (or fractal) wavelet-leader modeling that allowed the problem to be reformulated in a convex optimization framework by including a Total Variation penalization and (ii) investigating the potential of block-coordinate strategies in order to deal with the memory and computational cost induced by the minimization.

This presentation is part of Minisymposium “MS78 - Recent developments in variational image modeling
organized by: Sonia Tabti (Université de Caen, CNRS) , Rabin Julien (CNRS, Normandie Univ.) .

Nelly Pustelnik (CNRS, Laboratoire de Physique de l'ENS de Lyon)
Patrice Abry (CNRS, Laboratoire de Physique de l'ENS de Lyon)
Pascal Barbara (Laboratoire de Physique de l'ENS de Lyon)
image segmentation, nonlinear optimization, statistical inverse estimation methods