LiteBounD distills complementary semantic and boundary priors from multiple vision foundation models into compact segmentation backbones via dual-path and frequency-aware mechanisms, improving performance on both seen and unseen polyp datasets while preserving efficiency.
Selective feature aggrega- tion network with area-boundary constraints for polyp segmentation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Sharpening Lightweight Models for Generalized Polyp Segmentation: A Boundary Guided Distillation from Foundation Models
LiteBounD distills complementary semantic and boundary priors from multiple vision foundation models into compact segmentation backbones via dual-path and frequency-aware mechanisms, improving performance on both seen and unseen polyp datasets while preserving efficiency.