DISN: Deep Implicit Surface Network for High-quality Single-view 3D Reconstruction

TL;DR

DISN is a novel deep implicit surface network that reconstructs detailed 3D meshes from single-view images by combining global and local features to predict signed distance fields, capturing fine details like holes and thin structures.

Contribution

DISN introduces a new method that effectively captures detailed 3D structures from single images using combined global and local features for signed distance prediction.

Findings

Achieves state-of-the-art single-view 3D reconstruction results.

Successfully captures fine details such as holes and thin structures.

Performs well on both synthetic and real images.

Abstract

Reconstructing 3D shapes from single-view images has been a long-standing research problem. In this paper, we present DISN, a Deep Implicit Surface Network which can generate a high-quality detail-rich 3D mesh from an 2D image by predicting the underlying signed distance fields. In addition to utilizing global image features, DISN predicts the projected location for each 3D point on the 2D image, and extracts local features from the image feature maps. Combining global and local features significantly improves the accuracy of the signed distance field prediction, especially for the detail-rich areas. To the best of our knowledge, DISN is the first method that constantly captures details such as holes and thin structures present in 3D shapes from single-view images. DISN achieves the state-of-the-art single-view reconstruction performance on a variety of shape categories reconstructed from both synthetic and real images. Code is available at https://github.com/xharlie/DISN The supplementary can be found at https://xharlie.github.io/images/neurips_2019_supp.pdf