3D Reconstruction

Paper with code

3D RCNN (CVPR 2018)

3D-RCNN: Instance-level 3D Object Reconstruction via Render-and-Compare

Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild

Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild (CVPR 2020 (Best Paper Award))
Demo | Project Page | Video

Photo-Geometric Autoencoding: Our method is based on an autoencoder that factors each input image into depth, albedo, viewpoint and lighting (direction). These four components are combined to reconstruct the input image. The model is trained only using a reconstruction loss, without any external supervision.
Exploiting Symmetry: depth and albedo flipped horizontally to obtain 2 reconsturctions.
Probabilistic Modeling of Symmetry using Confidence Maps: adjust reconstruction loss for non-symmetric area.
depth + albedo -> 3d model (output)
depth + albedo + light + viewpoint -> render reconstructed image -> to compute loss