Please use this identifier to cite or link to this item: http://hdl.handle.net/10553/54429
Title: Symmetrical dense optical flow estimation with occlusions detection
Authors: Alvarez, Luis 
Deriche, Rachid
Papadopoulo, Théo
Sánchez, Javier 
UNESCO Clasification: 120601 Construcción de algoritmos
120602 Ecuaciones diferenciales
120326 Simulación
220990 Tratamiento digital. Imágenes
120304 Inteligencia artificial
Issue Date: 2007
Publisher: 0920-5691
Journal: International Journal of Computer Vision 
Abstract: Traditional techniques of dense optical flow estimation do not generally yield symmetrical solutions: the results will differ if they are applied between images I 1 and I 2 or between images I 2 and I 1. In this work, we present a method to recover a dense optical flow field map from two images, while explicitely taking into account the symmetry across the images as well as possible occlusions in the flow field. The idea is to consider both displacements vectors from I 1 to I 2 and I 2 to I 1 and to minimise an energy functional that explicitely encodes all those properties. This variational problem is then solved using the gradient flow defined by the Euler-Lagrange equations associated to the energy. To prove the importance of the concepts of symmetry and occlusions for optical flow computation, we have extended a classical approach to handle those. Experiments clearly show the added value of these properties to improve the accuracy of the computed flows. Figures appear in color in the online version of this paper. © 2007 Springer Science+Business Media, LLC.
URI: http://hdl.handle.net/10553/54429
ISSN: 0920-5691
DOI: 10.1007/s11263-007-0041-4
Source: International Journal of Computer Vision[ISSN 0920-5691],v. 75, p. 371-385
Appears in Collections:Artículos
Show full item record

SCOPUSTM   
Citations

97
checked on Jun 20, 2021

WEB OF SCIENCETM
Citations

76
checked on Jun 20, 2021

Page view(s)

44
checked on Jun 19, 2021

Google ScholarTM

Check

Altmetric


Share



Export metadata



Items in accedaCRIS are protected by copyright, with all rights reserved, unless otherwise indicated.