Reasoning 3D shapes from 2D images is an essential yet challenging task, especially when only single-view images are at our disposal. While an object can have a complicated shape, individual parts are usually close to geometric primitives and thus are easier to model. Furthermore, parts provide a mid-level representation that is robust to appearance variations across objects in a particular category. In this work, we tackle the problem of 3D part discovery from only 2D image collections. Instead of relying on manually annotated parts for supervision, we propose a self-supervised approach, latent part discovery (LPD). Our key insight is to learn a novel part shape prior that allows each part to fit an object shape faithfully while constrained to have simple geometry. Extensive experiments on the synthetic ShapeNet, PartNet, and real-world Pascal 3D+ datasets show that our method discovers consistent object parts and achieves favorable reconstruction accuracy compared to the existing methods with the same level of supervision.
@article{arxiv.2107.13629,
title = {Discovering 3D Parts from Image Collections},
author = {Chun-Han Yao and Wei-Chih Hung and Varun Jampani and Ming-Hsuan Yang},
journal= {arXiv preprint arXiv:2107.13629},
year = {2021}
}
Comments
Accepted by ICCV 2021. Project page: https://chhankyao.github.io/lpd/