ST-SAM combines self-training with entropy-based pseudo-label filtering and SAM prompt-based mutual correction to achieve strong camouflaged object detection with only 1 percent labeled data.
Segment Anything
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
-
ST-SAM: SAM-Driven Self-Training Framework for Semi-Supervised Camouflaged Object Detection
ST-SAM combines self-training with entropy-based pseudo-label filtering and SAM prompt-based mutual correction to achieve strong camouflaged object detection with only 1 percent labeled data.