A one-shot training regime with DINOv2-enriched point clouds and joint cross-attention predicts 3D object-to-object affordance maps that guide optimization-based robotic manipulation.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.RO 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
O$^3$Afford: One-Shot 3D Object-to-Object Affordance Grounding for Generalizable Robotic Manipulation
A one-shot training regime with DINOv2-enriched point clouds and joint cross-attention predicts 3D object-to-object affordance maps that guide optimization-based robotic manipulation.