Sobolev gradients and image interpolation

From MaRDI portal
Publication:2912283

DOI10.1137/11082453XzbMATH Open1253.65127arXiv1103.0720MaRDI QIDQ2912283FDOQ2912283


Authors: Ionut Danaila, Parimah Kazemi Edit this on Wikidata


Publication date: 14 September 2012

Published in: SIAM Journal on Imaging Sciences (Search for Journal in Brave)

Abstract: We present here a new image inpainting algorithm based on the Sobolev gradient method in conjunction with the Navier-Stokes model. The original model of Bertalmio et al is reformulated as a variational principle based on the minimization of a well chosen functional by a steepest descent method. This provides an alternative of the direct solving of a high-order partial differential equation and, consequently, allows to avoid complicated numerical schemes (min-mod limiters or anisotropic diffusion). We theoretically analyze our algorithm in an infinite dimensional setting using an evolution equation and obtain global existence and uniqueness results as well as the existence of an omega-limit. Using a finite difference implementation, we demonstrate using various examples that the Sobolev gradient flow, due to its smoothing and preconditioning properties, is an effective tool for use in the image inpainting problem.


Full work available at URL: https://arxiv.org/abs/1103.0720




Recommendations





Cited In (7)





This page was built for publication: Sobolev gradients and image interpolation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2912283)