Get Started
Home
Topics
Search
Library
Research questionHow can models segment arbitrary visual concepts precisely from one or a few annotated exemplars without task-specific training?The model must infer which pixels belong to a concept from only a small number of labeled examples. Foreground responses can be coarse or ambiguous, while background content can obscure complete object or part boundaries.
AI
Computer Vision
Image & Video Processing
Machine Learning
Latest papersRecent research connected to this question, newest first.FoRIS: Progressive Foreground Refinement for Training-Free In-Context SegmentationApplies to training-free in-context segmentation of semantic concepts, including objects and parts, using one-shot or five-shot annotated visual exemplars. The supplied evidence reports average mIoU improvements over existing approaches in these settings, but does not specify broader data, compute, or deployment constraints.research paper · Sep 3, 2026
Related questions
How can zero-shot 6D pose front-ends detect partially visible objects while rejecting similar distractors?How can fine-grained audio-visual segmentation learn new classes continually without semantic drift or co-occurrence confusion?How can text-promptable video segmentation track targets through disappearance while rejecting visually similar artifacts?How can plant point-cloud segmentation adapt across species and sensing conditions with few labeled examples?