MonoArt

3D models generated from photos are usually a single solid, with drawers and doors that don't open, useless for simulation. MonoArt reconstructs articulated 3D objects with movable parts from a single image, using progressive structural reasoning to find parts and their motions.

Output is ready for robot simulation and interactive apps.

Features

Articulation:Movable parts.

Reasoning:Step by step.

Single image:One photo.