articulation

FAMOS is a feed-forward model that predicts movable-part segmentation and joint parameters from sparse point clouds using a Multi-state Articulation Transformer and a procedural data generator, showing consistent improvements over baselines in experiments.