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

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

PDF Abstract NeurIPS 2019 PDF NeurIPS 2019 Abstract


Introduced in the Paper:


Used in the Paper:

ShapeNet ShapeNetCore
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Single-View 3D Reconstruction ShapeNetCore DISN 3DIoU 0.57 # 1
Single-View 3D Reconstruction ShapeNetCore OccNet 3DIoU 0.564 # 2
Single-View 3D Reconstruction ShapeNetCore IMNET 3DIoU 0.546 # 3
Single-View 3D Reconstruction ShapeNetCore 3DN 3DIoU 0.487 # 4
Single-View 3D Reconstruction ShapeNetCore Pxl2mesh 3DIoU 0.473 # 5
Single-View 3D Reconstruction ShapeNetCore AtlasNet 3DIoU 0,3 # 6


No methods listed for this paper. Add relevant methods here