A dual-branch state space model combining whole-image and windowed local scanning with attention fusion achieves 92.81% top-1 accuracy on the 27-class RSCD road surface dataset, ahead of the compared Mamba, Transformer, and CNN baselines.
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
-
RoadMamba: A Dual Branch Visual State Space Model for Road Surface Classification
A dual-branch state space model combining whole-image and windowed local scanning with attention fusion achieves 92.81% top-1 accuracy on the 27-class RSCD road surface dataset, ahead of the compared Mamba, Transformer, and CNN baselines.