Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo

Gottfried Graber, Thomas Pock, Stefano Soatto, Jonathan Balzer

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

We propose a method for dense three-dimensional surface reconstruction that leverages the strengths of shape-based approaches, by imposing regularization that respects
the geometry of the surface, and the strength of depthmap-based stereo, by avoiding costly computation of surface topology. The result is a near real-time variational reconstruction algorithm free of the staircasing artifacts that affect depth-map and plane-sweeping approaches. This is made possible by exploiting the gauge ambiguity to design
a novel representation of the regularizer that is linear in the parameters and hence amenable to be optimized with state-of-the-art primal-dual numerical schemes.
Original languageEnglish
Title of host publicationThe IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Publisher.
Pages511-520
Publication statusPublished - 2015
EventCVPR - Boston, USA, United States
Duration: 7 Jun 201512 Jun 2015

Conference

ConferenceCVPR
Abbreviated titleCVPR
CountryUnited States
CityBoston, USA
Period7/06/1512/06/15

Fingerprint

Surface reconstruction
Gages
Topology
Geometry

Fields of Expertise

  • Information, Communication & Computing

Treatment code (Nähere Zuordnung)

  • Application

Cite this

Graber, G., Pock, T., Soatto, S., & Balzer, J. (2015). Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 511-520). ..

Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo. / Graber, Gottfried; Pock, Thomas; Soatto, Stefano; Balzer, Jonathan.

The IEEE Conference on Computer Vision and Pattern Recognition (CVPR). ., 2015. p. 511-520.

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Graber, G, Pock, T, Soatto, S & Balzer, J 2015, Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo. in The IEEE Conference on Computer Vision and Pattern Recognition (CVPR). ., pp. 511-520, CVPR, Boston, USA, United States, 7/06/15.
Graber G, Pock T, Soatto S, Balzer J. Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR). . 2015. p. 511-520
Graber, Gottfried ; Pock, Thomas ; Soatto, Stefano ; Balzer, Jonathan. / Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR). ., 2015. pp. 511-520
@inproceedings{cd60c5c08ac24a41a9c9bb60fb2e75f2,
title = "Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo",
abstract = "We propose a method for dense three-dimensional surface reconstruction that leverages the strengths of shape-based approaches, by imposing regularization that respectsthe geometry of the surface, and the strength of depthmap-based stereo, by avoiding costly computation of surface topology. The result is a near real-time variational reconstruction algorithm free of the staircasing artifacts that affect depth-map and plane-sweeping approaches. This is made possible by exploiting the gauge ambiguity to designa novel representation of the regularizer that is linear in the parameters and hence amenable to be optimized with state-of-the-art primal-dual numerical schemes.",
author = "Gottfried Graber and Thomas Pock and Stefano Soatto and Jonathan Balzer",
year = "2015",
language = "English",
pages = "511--520",
booktitle = "The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)",
publisher = ".",

}

TY - GEN

T1 - Efficient Minimal-Surface Regularization of Perspective Depth Maps in Variational Stereo

AU - Graber, Gottfried

AU - Pock, Thomas

AU - Soatto, Stefano

AU - Balzer, Jonathan

PY - 2015

Y1 - 2015

N2 - We propose a method for dense three-dimensional surface reconstruction that leverages the strengths of shape-based approaches, by imposing regularization that respectsthe geometry of the surface, and the strength of depthmap-based stereo, by avoiding costly computation of surface topology. The result is a near real-time variational reconstruction algorithm free of the staircasing artifacts that affect depth-map and plane-sweeping approaches. This is made possible by exploiting the gauge ambiguity to designa novel representation of the regularizer that is linear in the parameters and hence amenable to be optimized with state-of-the-art primal-dual numerical schemes.

AB - We propose a method for dense three-dimensional surface reconstruction that leverages the strengths of shape-based approaches, by imposing regularization that respectsthe geometry of the surface, and the strength of depthmap-based stereo, by avoiding costly computation of surface topology. The result is a near real-time variational reconstruction algorithm free of the staircasing artifacts that affect depth-map and plane-sweeping approaches. This is made possible by exploiting the gauge ambiguity to designa novel representation of the regularizer that is linear in the parameters and hence amenable to be optimized with state-of-the-art primal-dual numerical schemes.

UR - http://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Graber_Efficient_Minimal-Surface_Regularization_2015_CVPR_paper.pdf

M3 - Conference contribution

SP - 511

EP - 520

BT - The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

PB - .

ER -