WeatherRemover is a lightweight all-in-one adverse weather removal model that uses channel-wise attention, linear spatial reduction, and gating in a multi-scale transformer-UNet to restore images efficiently across rain, snow, and fog.
End-to-end object detection with transformers,
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
-
WeatherRemover: All-in-one Adverse Weather Removal with Multi-scale Feature Map Compression
WeatherRemover is a lightweight all-in-one adverse weather removal model that uses channel-wise attention, linear spatial reduction, and gating in a multi-scale transformer-UNet to restore images efficiently across rain, snow, and fog.