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 3D reconstruction from a single 2D image (RGB), taken from the test set of ShapeNetCore dataset. We also provide two state-of-the-art methods using other 3D representations : 3D-R2N2 using voxels, and PointSetGen using point clouds.

• Homepage
• Single-View Reconstruction
• Auto Encoder
• Paper
• Code