arXiv Analytics

Sign in

arXiv:2003.03711 [cs.CV]AbstractReferencesReviewsResources

Meta3D: Single-View 3D Object Reconstruction from Shape Priors in Memory

Shuo Yang, Min Xu, Hongxun Yao

Published 2020-03-08Version 1

3D shape reconstruction from a single-view RGB image is an ill-posed problem due to the invisible parts of the object to be reconstructed. Most of the existing methods rely on large-scale data to obtain shape priors through tuning parameters of reconstruction models. These methods might not be able to deal with the cases with heavy object occlusions and noisy background since prior information can not be retained completely or applied efficiently. In this paper, we are the first to develop a memory-based meta-learning framework for single-view 3D reconstruction. A write controller is designed to extract shape-discriminative features from images and store image features and their corresponding volumes into external memory. A read controller is proposed to sequentially encode shape priors related to the input image and predict a shape-specific refiner. Experimental results demonstrate that our Meta3D outperforms state-of-the-art methods with a large margin through retaining shape priors explicitly, especially for the extremely difficult cases.

Related articles: Most relevant | Search more
arXiv:1411.2942 [cs.CV] (Published 2014-11-11)
3D Shape Reconstruction from 2D Landmarks: A Convex Formulation
arXiv:1707.06375 [cs.CV] (Published 2017-07-20)
3D Shape Reconstruction from Sketches via Multi-view Convolutional Networks
arXiv:1811.12016 [cs.CV] (Published 2018-11-29)
3D Shape Reconstruction from a Single 2D Image via 2D-3D Self-Consistency