Multinex：基于多先验 Retinex 的轻量低光照图像增强
Multinex: Lightweight Low-light Image Enhancement via Multi-prior Retinex
把 Retinex 残差形式与多先验（光照/颜色先验栈）结合，用极轻量网络只学亮度与反射率调整量而非重建整图，为边缘端低光照增强提供了一个可复用的分解-融合范式与 45K/0.7K 参数级的基线。
原文依据 an ultra-lightweight structured framework that integrates multiple fine-grained representations within a principled Retinex residual formulation
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论低光照图像增强（LLIE）旨在在严重光照退化条件下恢复自然的可见度、色彩保真度与结构细节。最优（SOTA）的 LLIE 技术往往依赖大模型与多阶段训练，限制了其在边缘部署中的实用性。此外，它们对单一颜色空间的依赖会带来不稳定性以及可见的曝光或颜色伪影。为解决这些问题，我们提出 Multinex，一个超轻量的结构化框架，它在有原则的 Retinex 残差形式中整合多种细粒度表示。该方法将图像分解为由不同解析表示导出的光照先验栈与颜色先验栈，并学习将这些表示融合为校正曝光所需的亮度与反射率调整量。通过将增强置于重建之上并利用轻量神经操作，Multinex 显著降低了计算开销，其轻量版（45K 参数）与纳米版（0.7K 参数）即为例证。大量基准测试表明，所有轻量变体均显著优于对应的轻量 SOTA 模型，并达到与重型模型相当的性能。论文页面见 https://albrateanu.github.io/multinex。
英文原文摘要 1277 字符
Low-light image enhancement (LLIE) aims to restore natural visibility, color fidelity, and structural detail under severe illumination degradation. State-of-the-art (SOTA) LLIE techniques often rely on large models and multi-stage training, limiting practicality for edge deployment. Moreover, their dependence on a single color space introduces instability and visible exposure or color artifacts.To address these, we propose Multinex, an ultra-lightweight structured framework that integrates multiple fine-grained representations within a principled Retinex residual formulation. It decomposes an image into illumination and color prior stacks derived from distinct analytic representations, and learns to fuse these representations into luminance and reflectance adjustments required to correct exposure. By prioritizing enhancement over reconstruction and exploiting lightweight neural operations, Multinex significantly reduces computational cost, exemplified by its lightweight (45K parameters) and nano (0.7K parameters) versions. Extensive benchmarks show that all lightweight variants significantly outperform their corresponding lightweight SOTA models, and reach comparable performance to heavy models. Paper page available at https://albrateanu.github.io/multinex.