SuperPatchMatch: an Algorithm for Robust Correspondences using Superpixel Patches
GIRAUD, Rémi
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Institut de Mathématiques de Bordeaux [IMB]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Institut de Mathématiques de Bordeaux [IMB]
TA, Vinh-Thong
Institut Polytechnique de Bordeaux [Bordeaux INP]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
See more >
Institut Polytechnique de Bordeaux [Bordeaux INP]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
GIRAUD, Rémi
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Institut de Mathématiques de Bordeaux [IMB]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Institut de Mathématiques de Bordeaux [IMB]
TA, Vinh-Thong
Institut Polytechnique de Bordeaux [Bordeaux INP]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
< Reduce
Institut Polytechnique de Bordeaux [Bordeaux INP]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Language
en
Article de revue
This item was published in
IEEE Transactions on Image Processing. 2017
Institute of Electrical and Electronics Engineers
English Abstract
Superpixels have become very popular in many computer vision applications. Nevertheless, they remain underexploited since the superpixel decomposition may produce irregular and non stable segmentation results due to the ...Read more >
Superpixels have become very popular in many computer vision applications. Nevertheless, they remain underexploited since the superpixel decomposition may produce irregular and non stable segmentation results due to the dependency to the image content. In this paper, we first introduce a novel structure, a superpixel-based patch, called SuperPatch. The proposed structure, based on superpixel neighborhood, leads to a robust descriptor since spatial information is naturally included. The generalization of the PatchMatch method to SuperPatches, named SuperPatchMatch, is introduced. Finally, we propose a framework to perform fast segmentation and labeling from an image database, and demonstrate the potential of our approach since we outperform, in terms of computational cost and accuracy, the results of state-of-the-art methods on both face labeling and medical image segmentation.Read less <
English Keywords
Patch-based method
PatchMatch
Labeling
Superpixels
Segmentation
ANR Project
Generalized Optimal Transport Models for Image processing - ANR-16-CE33-0010
Origin
Hal imported