AtlasNet: A Papier-Mâché Approach to Learning Mesh Synthesis
Thibault Groueix, Matthew Fisher, Vladimir G. Kim, Bryan C. Russell, Mathieu Aubry
Click on the images on the left panel to view different results of autoencoder's reconstruction of 3D meshes from 3D point clouds, taken from the test set of ShapeNetCore dataset. We show the ground truth mesh, and two reconstructions using two different versions of our algorithm.
Homepage
Single-View Reconstruction
Auto Encoder
Paper
Code
Selection tool
Ground Truth Mesh (Input)
AtlasNet (25 learned parameterizations) - Mesh
AtlasNet (1 sphere) - Mesh