A training-free pipeline that decomposes captions into phrase and word prompts and samples separate foreground and background points for SAM achieves state-of-the-art camouflaged object 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
-
Stepwise Decomposition and Dual-stream Focus: A Novel Approach for Training-free Camouflaged Object Segmentation
A training-free pipeline that decomposes captions into phrase and word prompts and samples separate foreground and background points for SAM achieves state-of-the-art camouflaged object segmentation.