A Spatial Regularization Approach for Vector Quantization

Abstract

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Iterative Multiplicative Filters for Data Labeling

“…Since the weights in the label assignment matrix are updated using only few of their local or nonlocal neighbors, our algorithm has a high potential for parallelization, which is not exploited in the implementation so far. In particular, a large number of labels does not increase the convergence speed drastically, which is the case for various TV-regularized variational methods as [8,14,17,16,36]. There are still a lot of open questions, which we want to examine in our future work, e.g.…”

“…It is the Bregman distance of the Shannon entropy and is related to entropy regularized Wasserstein distances, which have recently found applications in image processing, see [44] and the references therein. Setting the Riemannian gradient of the objective in (16) to zero using (17), and dividing by 2 we get…”

Image processing for materials characterization: Issues, challenges and opportunities

“…For illustrating and to widen the points of view, some of these challenges can be undertaken by employing active sub-fields of image processing, e.g. restoration and inverse problems [15]; segmentation [16,17,18]; texture analysis [19,20]; multiscale and directional features extraction methods [21]; color and multispectral processing [12]; stochastic models [22,23].…”

Segmentation of images with separating layers by fuzzy c-means and convex optimization

“…their weighted version or the Kullback-Leibler divergence, see [6]. As regularization term Ψ(l) a discrete version of the Rudin-Osher-Fatemi T V -functional [26] is a good candidate since it enforces 'smooth' boundaries between the labeled regions.…”

“…For λ = 0, the optimization becomes much harder. An approach via discrete optimization for the anisotropic T V -functional which involves graph-cut based algorithms was proposed in [6].…”