RetinexDualV2: Physically-Grounded Dual Retinex for Generalized UHD Image Restoration

TL;DR

RetinexDualV2 introduces a physically grounded dual-branch framework for UHD image restoration, leveraging degradation-aware priors and a novel attention mechanism to handle diverse degradations effectively.

Contribution

It presents a unified architecture with a physical grounding module and a new attention mechanism, enabling robust, generalizable UHD image restoration without task-specific modifications.

Findings

Achieved 4th place in NTIRE 2026 Raindrop Removal Challenge.

Secured 5th place in JNLLIE Low-light Enhancement Challenge.

Demonstrated state-of-the-art performance and efficiency.

Abstract

We propose RetinexDualV2, a unified, physically grounded dual-branch framework for diverse Ultra-High-Definition (UHD) image restoration. Unlike generic models, our method employs a Task-Specific Physical Grounding Module (TS-PGM) to extract degradation-aware priors (e.g., rain masks and dark channels). These explicitly guide a Retinex decomposition network via a novel Physical-Conditioned Multi-head Self-Attention (PC-MSA) mechanism, enabling robust reflection and illumination correction. This physical conditioning allows a single architecture to handle various complex degradations seamlessly, without task-specific structural modifications. RetinexDualV2 demonstrates exceptional generalizability, securing 4th place in the NTIRE 2026 Day and Night Raindrop Removal Challenge and 5th place in the Joint Noise Low-light Enhancement (JNLLIE) Challenge. Extensive experiments confirm the state-of-the-art performance and efficiency of our physically motivated approach. Code is available at https://github.com/ErrorLogic1211/RetinexDual/tree/master/RetinexDualV2