Functionality, Articulation, and Interaction Workshop
On modeling and generating 3D objects that constitute interactive environments.
On modeling and generating 3D objects that constitute interactive environments.
Recent advances in 3D generative AI have achieved impressive geometric and textural fidelity. Yet three critical dimensions remain overlooked: functionality, articulation, and interaction. Generated 3D content must not only look realistic — it must move correctly, respond to physical forces, and afford meaningful engagement with agents and environments.
The FAI workshop brings together researchers across computer vision, graphics, robotics, and machine learning to unify work on functional understanding, articulated reconstruction, part decomposition, and physical grounding. We position functionality as a first-class concept across the full pipeline — perceived from visual observations, represented through explicit structure, and realized through physically grounded generation.
We welcome submissions on modeling and generation of functional and interactable 3D assets. Both short papers (7 pages) and full papers (14 pages) are welcome, in ECCV format (excluding references). All accepted submissions will be presented as posters. Papers will not be included in the official proceedings.
