AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

PR-MaGIC: Prompt Refinement Via Mask Decoder Gradient Flow For In-Context Segmentation

Minjae Lee, Sungwoo Hur, Soojin Hwang, Won Hwa Kim

CVPR 2026 · Detection and segmentation › Open-set and few-shot detection · approach: Prompted SAM segmentation · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add PR-MaGIC gradient-flow refinement with top-1 selectionPerSAM-F · FSS-1000mIoU: 85.85 -> 87.26Improved
add PR-MaGIC with top-1 selectionPerSAM-F · COCO-20imIoU: 53.36 -> 55.96Improved
add PR-MaGIC with top-1 selectionMatcher (1-shot) · FSS-1000mIoU: 92.08 -> 92.06No clear effect
add PR-MaGIC with top-1 selectionMatcher (1-shot) · DIS5KmIoU: 46.65 -> 55.08Improved
add PR-MaGIC with top-1 selectionMatcher (5-shot) · COCO-20imIoU: 67.69 -> 70.74Improved
add PR-MaGIC with top-1 selectionMatcher (1-shot) · PACO-PartmIoU: 50.27 -> 54.08Improved
use oracle selection instead of top-1 similarity selection (ablation)PR-MaGIC on PerSAM-F · FSS-1000mIoU: 87.26 -> 91.88Improved
use a fixed number of refinement iterations instead of top-1 selection (ablation)PR-MaGIC on Matcher (1-shot) · COCO-20imIoU gain: -0.15, 0.21, 0.07, 0.31, 0.66Mixed
use fixed 5 iterations instead of top-1 selection (ablation)PR-MaGIC on PerSAM-F (1-shot) · COCO-20imIoU gain: -0.43 vs +2.60 (Top-1)Hurt
increase step size eta to 1e-2 (ablation)PR-MaGIC · COCO-20i (100 train samples)mIoU: rapid early gains but degrades with more iterationsMixed
decrease step size eta to 1e-4 or 1e-5 (ablation)PR-MaGIC · COCO-20i (100 train samples)mIoU: saturates below best achievable mIoUHurt
larger step size eta (ablation)PR-MaGIC on PerSAM-F · FSS-1000 (9 classes)mIoU: 70.88 > 67.89 > 64.78Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy