SAOR: Single-View Articulated Object Reconstruction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Aygün, Mehmet, Mac Aodha, Oisin
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914742659973120
author Aygün, Mehmet
Mac Aodha, Oisin
author_facet Aygün, Mehmet
Mac Aodha, Oisin
contents We introduce SAOR, a novel approach for estimating the 3D shape, texture, and viewpoint of an articulated object from a single image captured in the wild. Unlike prior approaches that rely on pre-defined category-specific 3D templates or tailored 3D skeletons, SAOR learns to articulate shapes from single-view image collections with a skeleton-free part-based model without requiring any 3D object shape priors. To prevent ill-posed solutions, we propose a cross-instance consistency loss that exploits disentangled object shape deformation and articulation. This is helped by a new silhouette-based sampling mechanism to enhance viewpoint diversity during training. Our method only requires estimated object silhouettes and relative depth maps from off-the-shelf pre-trained networks during training. At inference time, given a single-view image, it efficiently outputs an explicit mesh representation. We obtain improved qualitative and quantitative results on challenging quadruped animals compared to relevant existing work.
format Preprint
id arxiv_https___arxiv_org_abs_2303_13514
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle SAOR: Single-View Articulated Object Reconstruction
Aygün, Mehmet
Mac Aodha, Oisin
Computer Vision and Pattern Recognition
We introduce SAOR, a novel approach for estimating the 3D shape, texture, and viewpoint of an articulated object from a single image captured in the wild. Unlike prior approaches that rely on pre-defined category-specific 3D templates or tailored 3D skeletons, SAOR learns to articulate shapes from single-view image collections with a skeleton-free part-based model without requiring any 3D object shape priors. To prevent ill-posed solutions, we propose a cross-instance consistency loss that exploits disentangled object shape deformation and articulation. This is helped by a new silhouette-based sampling mechanism to enhance viewpoint diversity during training. Our method only requires estimated object silhouettes and relative depth maps from off-the-shelf pre-trained networks during training. At inference time, given a single-view image, it efficiently outputs an explicit mesh representation. We obtain improved qualitative and quantitative results on challenging quadruped animals compared to relevant existing work.
title SAOR: Single-View Articulated Object Reconstruction
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2303.13514