Reinforcement learning can teach an LMM to prompt SAM2 for segmentation, achieving competitive camouflaged and salient object detection and zero-shot referring segmentation.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Seg-R1: Segmentation Can Be Surprisingly Simple with Reinforcement Learning
Reinforcement learning can teach an LMM to prompt SAM2 for segmentation, achieving competitive camouflaged and salient object detection and zero-shot referring segmentation.