Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View

Albert Pumarola, Antonio Agudo, Lorenzo Porzi, Alberto Sanfeliu, Vincent Lepetit, Francesc Moreno-Noguer

Research output: Contribution to journalArticle

Abstract

We propose a method for predicting the 3D shape of a deformable surface from a single view. By contrast with previous approaches, we do not need a pre-registered template of the surface, and our method is robust to the lack of texture and partial occlusions. At the core of our approach is a {\it geometry-aware} deep architecture that tackles the problem as usually done in analytic solutions: first perform 2D detection of the mesh and then estimate a 3D shape that is geometrically consistent with the image. We train this architecture in an end-to-end manner using a large dataset of synthetic renderings of shapes under different levels of deformation, material properties, textures and lighting conditions. We evaluate our approach on a test split of this dataset and available real benchmarks, consistently improving state-of-the-art solutions with a significantly lower computational time.

Keywords

  • cs.CV

Cite this

Pumarola, A., Agudo, A., Porzi, L., Sanfeliu, A., Lepetit, V., & Moreno-Noguer, F. (2018). Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View. arXiv.org e-Print archive.

Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View. / Pumarola, Albert; Agudo, Antonio; Porzi, Lorenzo; Sanfeliu, Alberto; Lepetit, Vincent; Moreno-Noguer, Francesc.

In: arXiv.org e-Print archive, 27.09.2018.

Research output: Contribution to journalArticle

Pumarola, A, Agudo, A, Porzi, L, Sanfeliu, A, Lepetit, V & Moreno-Noguer, F 2018, 'Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View' arXiv.org e-Print archive.
Pumarola A, Agudo A, Porzi L, Sanfeliu A, Lepetit V, Moreno-Noguer F. Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View. arXiv.org e-Print archive. 2018 Sep 27.
Pumarola, Albert ; Agudo, Antonio ; Porzi, Lorenzo ; Sanfeliu, Alberto ; Lepetit, Vincent ; Moreno-Noguer, Francesc. / Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View. In: arXiv.org e-Print archive. 2018
@article{461117b5d7f5410db330a976e0ff4264,
title = "Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View",
abstract = "We propose a method for predicting the 3D shape of a deformable surface from a single view. By contrast with previous approaches, we do not need a pre-registered template of the surface, and our method is robust to the lack of texture and partial occlusions. At the core of our approach is a {\it geometry-aware} deep architecture that tackles the problem as usually done in analytic solutions: first perform 2D detection of the mesh and then estimate a 3D shape that is geometrically consistent with the image. We train this architecture in an end-to-end manner using a large dataset of synthetic renderings of shapes under different levels of deformation, material properties, textures and lighting conditions. We evaluate our approach on a test split of this dataset and available real benchmarks, consistently improving state-of-the-art solutions with a significantly lower computational time.",
keywords = "cs.CV",
author = "Albert Pumarola and Antonio Agudo and Lorenzo Porzi and Alberto Sanfeliu and Vincent Lepetit and Francesc Moreno-Noguer",
note = "Accepted at CVPR 2018",
year = "2018",
month = "9",
day = "27",
language = "English",
journal = "arXiv.org e-Print archive",
publisher = "Cornell University Library",

}

TY - JOUR

T1 - Geometry-Aware Network for Non-Rigid Shape Prediction from a Single View

AU - Pumarola,Albert

AU - Agudo,Antonio

AU - Porzi,Lorenzo

AU - Sanfeliu,Alberto

AU - Lepetit,Vincent

AU - Moreno-Noguer,Francesc

N1 - Accepted at CVPR 2018

PY - 2018/9/27

Y1 - 2018/9/27

N2 - We propose a method for predicting the 3D shape of a deformable surface from a single view. By contrast with previous approaches, we do not need a pre-registered template of the surface, and our method is robust to the lack of texture and partial occlusions. At the core of our approach is a {\it geometry-aware} deep architecture that tackles the problem as usually done in analytic solutions: first perform 2D detection of the mesh and then estimate a 3D shape that is geometrically consistent with the image. We train this architecture in an end-to-end manner using a large dataset of synthetic renderings of shapes under different levels of deformation, material properties, textures and lighting conditions. We evaluate our approach on a test split of this dataset and available real benchmarks, consistently improving state-of-the-art solutions with a significantly lower computational time.

AB - We propose a method for predicting the 3D shape of a deformable surface from a single view. By contrast with previous approaches, we do not need a pre-registered template of the surface, and our method is robust to the lack of texture and partial occlusions. At the core of our approach is a {\it geometry-aware} deep architecture that tackles the problem as usually done in analytic solutions: first perform 2D detection of the mesh and then estimate a 3D shape that is geometrically consistent with the image. We train this architecture in an end-to-end manner using a large dataset of synthetic renderings of shapes under different levels of deformation, material properties, textures and lighting conditions. We evaluate our approach on a test split of this dataset and available real benchmarks, consistently improving state-of-the-art solutions with a significantly lower computational time.

KW - cs.CV

M3 - Article

JO - arXiv.org e-Print archive

T2 - arXiv.org e-Print archive

JF - arXiv.org e-Print archive

ER -