X-Restormer++ extends a dual-attention transformer with two-stage training on large and small datasets, a gradient-guided edge-aware loss, and weighted dual-model ensembling to achieve first place in an all-weather image restoration challenge.
Foundir: Unleashing million-scale training data to ad- vance foundation models for image restoration
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
-
X-Restormer++: 1st Place Solution for the UG2+ CVPR 2026 All-Weather Restoration Challenge
X-Restormer++ extends a dual-attention transformer with two-stage training on large and small datasets, a gradient-guided edge-aware loss, and weighted dual-model ensembling to achieve first place in an all-weather image restoration challenge.