PROXIMAL SPLITTING METHODS FOR DEPTH ESTIMATION

Stereoscopic and 3-D Coding

Presented by: Mireille El Gheche, Author(s): Mireille El Gheche, Jean-Christophe Pesquet, Université Paris-Est Marne la Vallée, France; Joumana Farah, Holy-Spirit University of Kaslik, Lebanon; Mounir Kaaniche, Béatrice Pesquet-Popescu, Télécom ParisTech, France

Stereo matching is an active area of research in image processing. In a recent work, a convex programming approach has been proposed in order to generate a dense disparity field. In this paper, we address the same estimation problem and propose to solve it by using a more general convex optimization framework based on proximal methods. More precisely, unlike previous works where the criterion must satisfy some restrictive conditions in order to be able to solve numerically the minimization problem, this work offers a great flexibility in the choice of the involved criterion. Experimental results carried out on real stereo pairs demonstrate the good performance of the proposed parallel proximal algorithm.

You need the Flash Player.

Share:

Download subtitles | Enlarge video

1. slide