Cahiers du CEREMADE

Unité Mixte de Recherche du C.N.R.S. N°7534
Abstract : This paper presents a generative model for textures that uses a local sparse description of the image content. This model enforces the sparsity of the expansion of local texture patches on adapted atomic elements. We focus on two approaches to define these atoms and to compute the corresponding sparse decomposition. The first one is inspired by recently proposed algorithms for non-local denoising. The second one imposes strict sparsity and optimizes the dictionary in order to sparsify the set of patches in the exemplar texture. These two methods define a texture ensemble that captures the patterns of the input exemplar. An iterative algorithm enables a fast texture synthesis that draws a random image from this texture ensemble. Similar algorithms can be used to perform texture inpainting and texture segmentation.
Sparse Modeling of Textures
PEYRE Gabriel
Université de PARIS - DAUPHINE
Place du Maréchal de Lattre De Tassigny - 75775 PARIS CEDEX 16 - FRANCE
Téléphone : +33 (0)1 44-05-49-23 - fax : +33 (0)1 44-05-45-99