GLARE: Low Light Image Enhancement via Generative Latent Feature based Codebook Retrieval

TL;DR

GLARE introduces a novel low-light image enhancement method using generative latent features and codebook retrieval, effectively improving image quality in extremely low-light conditions and benefiting high-level vision tasks.

Contribution

The paper proposes a new LLIE network combining codebook retrieval with generative latent features and an invertible flow, addressing the ill-posed nature of low-light enhancement.

Findings

Outperforms existing methods on benchmark datasets

Enhances low-light image quality with realistic details

Improves low-light object detection performance

Abstract

Most existing Low-light Image Enhancement (LLIE) methods either directly map Low-Light (LL) to Normal-Light (NL) images or use semantic or illumination maps as guides. However, the ill-posed nature of LLIE and the difficulty of semantic retrieval from impaired inputs limit these methods, especially in extremely low-light conditions. To address this issue, we present a new LLIE network via Generative LAtent feature based codebook REtrieval (GLARE), in which the codebook prior is derived from undegraded NL images using a Vector Quantization (VQ) strategy. More importantly, we develop a generative Invertible Latent Normalizing Flow (I-LNF) module to align the LL feature distribution to NL latent representations, guaranteeing the correct code retrieval in the codebook. In addition, a novel Adaptive Feature Transformation (AFT) module, featuring an adjustable function for users and comprising an Adaptive Mix-up Block (AMB) along with a dual-decoder architecture, is devised to further enhance fidelity while preserving the realistic details provided by codebook prior. Extensive experiments confirm the superior performance of GLARE on various benchmark datasets and real-world data. Its effectiveness as a preprocessing tool in low-light object detection tasks further validates GLARE for high-level vision applications. Code is released at https://github.com/LowLevelAI/GLARE.