CANDLE: Illumination-Invariant Semantic Priors for Color Ambient Lighting Normalization

2026-04-03

AI Summary

This work addresses the challenging problem of intrinsic color recovery under multicolored environmental illumination, where severe color casts, highlight saturation, and material-dependent reflectance significantly degrade performance. To tackle this, we introduce DINOv2 features—leveraging their self-supervised semantic richness—into illumination normalization for the first time, proposing a novel D.O.G. (Deep-layer Optimized Guidance) multi-level feature fusion strategy that adaptively injects semantic priors via a full-layer guidance mechanism. Furthermore, we design a joint color-frequency optimization module, termed BFACG+SFFB, to effectively suppress chromatic collapse and detail contamination. Evaluated on the CL3AN dataset, our method achieves a 1.22 dB PSNR improvement and secured third place in the colored illumination track and second place in white-light fidelity (lowest FID) at the NTIRE 2026 ALN Challenge, demonstrating strong generalization and state-of-the-art performance under complex lighting conditions.

Abstract

Color ambient lighting normalization under multi-colored illumination is challenging due to severe chromatic shifts, highlight saturation, and material-dependent reflectance. Existing geometric and low-level priors are insufficient for recovering object-intrinsic color when illumination-induced chromatic bias dominates. We observe that DINOv3's self-supervised features remain highly consistent between colored-light inputs and ambient-lit ground truth, motivating their use as illumination-robust semantic priors. We propose CANDLE (Color Ambient Normalization with DINO Layer Enhancement), which introduces DINO Omni-layer Guidance (D.O.G.) to adaptively inject multi-layer DINOv3 features into successive encoder stages, and a color-frequency refinement design (BFACG + SFFB) to suppress decoder-side chromatic collapse and detail contamination. Experiments on CL3AN show a +1.22 dB PSNR gain over the strongest prior method. CANDLE achieves 3rd place on the NTIRE 2026 ALN Color Lighting Challenge and 2nd place in fidelity on the White Lighting track with the lowest FID, confirming strong generalization across both chromatic and luminance-dominant illumination conditions. Code is available at https://github.com/ron941/CANDLE.

Problem

• color ambient lighting normalization
• chromatic shift
• illumination-invariant
• intrinsic color recovery
• multi-colored illumination

Innovation

• illumination-invariant priors
• DINOv3 features
• color ambient lighting normalization
• multi-layer feature injection
• chromatic collapse suppression