One-Shot Transfer of Affordance Regions? AffCorrs!
This enables future applications like learning affordances via imitation and assisted teleoperation, though it's an incremental improvement over existing methods.
The paper tackles one-shot visual search of object parts by segmenting semantically corresponding affordance regions in target scenes using only a single reference image with annotations. The proposed AffCorrs model achieves this through unsupervised combination of pre-trained DINO-ViT descriptors and cyclic correspondences for both intra- and inter-class segmentation.
In this work, we tackle one-shot visual search of object parts. Given a single reference image of an object with annotated affordance regions, we segment semantically corresponding parts within a target scene. We propose AffCorrs, an unsupervised model that combines the properties of pre-trained DINO-ViT's image descriptors and cyclic correspondences. We use AffCorrs to find corresponding affordances both for intra- and inter-class one-shot part segmentation. This task is more difficult than supervised alternatives, but enables future work such as learning affordances via imitation and assisted teleoperation.