Internally Referenced Low-Light Enhancement

2026-05-27

AI Summary

This work addresses the challenge of self-supervised low-light image enhancement, where the absence of external ground truth makes it difficult to disentangle illumination, texture, and noise. To overcome this, the authors propose an intrinsic-reference-driven enhancement framework that constructs dual internal references—physical and structural—from a single degraded image. Specifically, local exposure simulation generates a low-frequency pseudo-ground truth to guide global illumination estimation, while spatial-spectral dual-domain constraints preserve structural details. Key innovations include an illumination-alignment-aware loss, a translation-invariant spectral correlation loss, and a gain-adaptive feature modulation mechanism (GAFM), which collectively enable effective decoupling of illumination, texture, and noise. The method achieves state-of-the-art performance across multiple benchmarks, significantly improving both denoising capability and texture fidelity.

Abstract

Self-supervised low-light image enhancement (LLIE) is highly appealing as it eliminates the reliance on external paired data. However, the lack of external references causes networks to struggle with decoupling entangled illumination, delicate textures, and amplified noise. To resolve this challenge, we propose an Internally Referenced LLIE framework that extracts reliable physical and structural references from the degraded input image itself. First, we introduce a local exposure-simulated scheme to extract a low-frequency pseudo ground-truth. This serves as an internal physical reference to guide global illumination estimation and correct color casts. Second, we propose a dual-domain preservation strategy with spatial and spectral constraints to construct internal structural references. Specifically, an Illumination-Aligned Perceptual loss preserves global structures under illumination shifts, while a Shift-Invariant Spectral Correlation loss captures fine-grained local structures and suppresses high-frequency noise. Finally, we propose a Gain-Adaptive Feature Modulation (GAFM) mechanism to address highly spatially-variant residual noise. By transforming the self-estimated illumination map into an internal spatial gain prior, GAFM dynamically guides a blind-spot network for spatially-aware denoising. Extensive experiments demonstrate that our method achieves state-of-the-art performance, delivering superior noise suppression and textural fidelity. Code will be publicly released at https://visonj.github.io/IRLE/.

Problem

• low-light image enhancement
• self-supervised learning
• illumination decoupling
• noise amplification
• texture preservation

Innovation

• self-supervised
• low-light enhancement
• internal reference
• dual-domain preservation
• gain-adaptive modulation