USAM trains small MLPs on SAM's mask and IoU tokens to estimate predictive, prompt, task, and model uncertainty, achieving strong selective-correction results at negligible computational overhead.
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
-
UncertainSAM: Fast and Efficient Uncertainty Quantification of the Segment Anything Model
USAM trains small MLPs on SAM's mask and IoU tokens to estimate predictive, prompt, task, and model uncertainty, achieving strong selective-correction results at negligible computational overhead.