Winner of CVPR2026 NTIRE Challenge on Image Shadow Removal: Semantic and Geometric Guidance for Shadow Removal via Cascaded Refinement

TL;DR

This paper introduces a three-stage shadow removal pipeline that leverages semantic and geometric cues, achieving top performance in the CVPR2026 NTIRE challenge.

Contribution

The authors propose a novel multi-stage refinement approach using semantic guidance and geometric cues, with a contraction-constrained objective for stable optimization.

Findings

Achieved 26.680 PSNR on the WSRD+ 2026 test set, ranking first.

Validated the model's effectiveness on ISTD+ and UAV-SC+ datasets.

Won the NTIRE 2026 Image Shadow Removal Challenge.

Abstract

We present a three-stage progressive shadow-removal pipeline for the CVPR2026 NTIRE WSRD+ challenge. Built on OmniSR, our method treats deshadowing as iterative direct refinement, where later stages correct residual artefacts left by earlier predictions. The model combines RGB appearance with frozen DINOv2 semantic guidance and geometric cues from monocular depth and surface normals, reused across all stages. To stabilise multi-stage optimisation, we introduce a contraction-constrained objective that encourages non-increasing reconstruction error across the cascade. A staged training pipeline transfers from earlier WSRD pretraining to WSRD+ supervision and final WSRD+ 2026 adaptation with cosine-annealed checkpoint ensembling. On the official WSRD+ 2026 hidden test set, our final ensemble achieved 26.680 PSNR, 0.8740 SSIM, 0.0578 LPIPS, and 26.135 FID, ranked first overall, and won the NTIRE 2026 Image Shadow Removal Challenge. The strong performance of the proposed model is further validated on the ISTD+ and UAV-SC+ datasets.