Specular-to-Diffuse Translation for Multi-view Reconstruction

Shihao Wu, Hui Huang*, Tiziano Portenier, Matan Sela, Daniel Cohen-Or, Ron Kimmel, Matthias Zwicker

*Corresponding author for this work

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

Abstract

Most multi-view 3D reconstruction algorithms, especially when shape-from-shading cues are used, assume that object appearance is predominantly diffuse. To alleviate this restriction, we introduce S2Dnet, a generative adversarial network for transferring multiple views of objects with specular reflection into diffuse ones, so that multi-view reconstruction methods can be applied more effectively. Our network extends unsupervised image-to-image translation to multi-view “specular to diffuse” translation. To preserve object appearance across multiple views, we introduce a Multi-View Coherence loss (MVC) that evaluates the similarity and faithfulness of local patches after the view-transformation. In addition, we carefully design and generate a large synthetic training data set using physically-based rendering. During testing, our network takes only the raw glossy images as input, without extra information such as segmentation masks or lighting estimation. Results demonstrate that multi-view reconstruction can be significantly improved using the images filtered by our network.

Original languageEnglish
Title of host publicationComputer Vision – ECCV 2018 - 15th European Conference, 2018, Proceedings
EditorsVittorio Ferrari, Cristian Sminchisescu, Yair Weiss, Martial Hebert
PublisherSpringer Verlag
Pages193-211
Number of pages19
ISBN (Print)9783030012243
DOIs
StatePublished - 2018
Event15th European Conference on Computer Vision, ECCV 2018 - Munich, Germany
Duration: 8 Sep 201814 Sep 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11208 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th European Conference on Computer Vision, ECCV 2018
Country/TerritoryGermany
CityMunich
Period8/09/1814/09/18

Keywords

  • Generative adversarial network
  • Image translation
  • Multi-view coherence
  • Multi-view reconstruction
  • Specular-to-diffuse

Fingerprint

Dive into the research topics of 'Specular-to-Diffuse Translation for Multi-view Reconstruction'. Together they form a unique fingerprint.

Cite this