SegMAN combines local neighborhood attention and Mamba state space scanning in a hybrid encoder-decoder to achieve strong semantic segmentation results on ADE20K, Cityscapes, and COCO-Stuff.
Per- pixel classification is not all you need for semantic segmen- tation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
SegMAN: Omni-scale Context Modeling with State Space Models and Local Attention for Semantic Segmentation
SegMAN combines local neighborhood attention and Mamba state space scanning in a hybrid encoder-decoder to achieve strong semantic segmentation results on ADE20K, Cityscapes, and COCO-Stuff.