M2Retinexformer: Multi-Modal Retinexformer for Low-Light Image Enhancement

TL;DR

M2Retinexformer is a multi-modal deep learning framework that enhances low-light images by integrating depth, luminance, and semantic cues through a progressive, attention-based fusion process.

Contribution

It introduces a novel multi-modal Retinex-based model that incorporates auxiliary cues and adaptive gating for improved low-light image enhancement.

Findings

Outperforms state-of-the-art methods on multiple benchmarks.

Effectively fuses multi-scale modalities via cross-attention and gating.

Achieves better noise reduction and color fidelity in low-light images.

Abstract

Low-light image enhancement is challenging due to complex degradations, including amplified noise, artifacts, and color distortion. While Retinex-based deep learning methods have achieved promising results, they primarily rely on single-modality RGB information. We propose M2Retinexformer (Multi-Modal Retinexformer), a novel framework that extends Retinexformer by incorporating depth cues, luminance priors, and semantic features within a progressive refinement pipeline. Depth provides geometric context that is invariant to lighting variations, while luminance and semantic features offer explicit guidance on brightness distribution and scene understanding. Modalities are extracted at multiple scales and fused through cross-attention, with adaptive gating dynamically balancing illumination-guided self-attention and cross-attention based on the reliability of auxiliary cues. Evaluations on the LOL, SID, SMID, and SDSD benchmarks demonstrate overall improvements over Retinexformer and recent state-of-the-art methods. Code and pretrained weights are available at https://github.com/YoussefAboelwafa/M2Retinexformer