PDSSNet reports state-of-the-art mIoU of 84.68, 87.55, and 56.10 on the Vaihingen, Potsdam, and LoveDA remote sensing datasets by combining GT-derived prototypes, a Mamba-style semantic-structure module, and a similarity-driven step-size mechanism.
Dgnet: Distribution guided efficient learning for oil spill image segmentation,
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
-
Prototype-Driven Structure Synergy Network for Remote Sensing Images Segmentation
PDSSNet reports state-of-the-art mIoU of 84.68, 87.55, and 56.10 on the Vaihingen, Potsdam, and LoveDA remote sensing datasets by combining GT-derived prototypes, a Mamba-style semantic-structure module, and a similarity-driven step-size mechanism.