Retinex model is an effective tool for low-light image enhancement. It assumes that observed images can be decomposed into the reflectance and illumination. Most existing Retinex-based methods have carefully designed hand-crafted constraints and parameters for this highly ill-posed decomposition, which may be limited by model capacity when applied in various scenes. In this paper, we collect a LOw-Light dataset (LOL) containing low/normal-light image pairs and propose a deep Retinex-Net learned on this dataset, including a Decom-Net for decomposition and an Enhance-Net for illumination adjustment. In the training process for Decom-Net, there is no ground truth of decomposed reflectance and illumination. The network is learned with only key constraints including the consistent reflectance shared by paired low/normal-light images, and the smoothness of illumination. Based on the decomposition, subsequent lightness enhancement is conducted on illumination by an enhancement network called Enhance-Net, and for joint denoising there is a denoising operation on reflectance. The Retinex-Net is end-to-end trainable, so that the learned decomposition is by nature good for lightness adjustment. Extensive experiments demonstrate that our method not only achieves visually pleasing quality for low-light enhancement but also provides a good representation of image decomposition.
IllumFlow: Illumination-Adaptive Low-Light Enhancement via Conditional Rectified Flow and Retinex Decomposition
School of Mathematics and Statistics, Xidian University(西安电子科技大学数学与统计学学院);School of Mathematics and Statistics, Xi'an Jiaotong University(西安交通大学数学与统计学学院)
Low-light Image Enhancement with Retinex Decomposition in Latent Space
School of Automation, Hangzhou Dianzi University(杭州电子科技大学自动化学院);College of Artificial Intelligence, Jiaxing University(嘉兴大学人工智能学院);School of Artificial Intelligence at Dalian University of Technology(大连理工大学人工智能学院);Department of Computer Science at City University of Hong Kong(香港城市大学计算机科学系)
DI-Retinex: Digital-Imaging Retinex Theory for Low-Light Image Enhancement
Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所);University of Chinese Academy of Sciences(中国科学院大学);The School of Cyber Security, University of Chinese Academy of Sciences(中国科学院大学网络空间安全学院);The School of Cyber Science and Technology, Shenzhen Campus of Sun Yat-sen University(中山大学深圳校区网络空间科学与技术学院);Huawei Noah’s Ark Lab(华为诺亚方舟实验室)
Pre-processing Image using Brightening, CLAHE and RETINEX
Centre of Digital Technologies Education Research (CODER)(数字技术教育研究中心(CODER));ICT, School of Technology, Environments & Design(技术、环境与设计学院信息与通信技术(ICT));University of Tasmania(塔斯马尼亚大学)
Nankai University(南开大学);Taishan University(泰山学院);Tianjin University(天津大学);Inception Institute of Artificial Intelligence(起源人工智能研究院);Tsinghua Shenzhen International Graduate School(清华大学深圳国际研究生院);Shenzhen Institute of Future Media Technology(深圳未来媒体技术研究院);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)
Retinex Meets Language: A Physics-Semantics-Guided Underwater Image Enhancement Network
State Key Laboratory of Physical Oceanography and the Faculty of Information Science and Engineering, Ocean University of China(物理海洋学国家重点实验室和中国海洋大学信息科学与工程学院);School of Cyber Science and Technology, Shenzhen Campus of Sun Yat-sen university(中山大学深圳校区计算机科学与技术学院)
Retinex-RAWMamba: Bridging Demosaicing and Denoising for Low-Light RAW Image Enhancement
Institute of Advanced Technology, University of Science and Technology of China(科学技术大学先进技术研究院);School of Computer and Artificial Intelligence, Beijing Technology and Business University(北京科技与商务大学计算机与人工智能学院);Institute of Artificial Intelligence, Hefei Comprehensive National Science Center(合肥综合性国家科学中心人工智能研究院);University of Science and Technology of China(科学技术大学);School of Automation, Northwestern Polytechnical University(西北工业大学自动化学院)
Deep Bilateral Retinex for Low-Light Image Enhancement
South China University of Technology(华南理工大学);Peng Cheng Laboratory(鹏城实验室);Tencent AI Lab(腾讯AI实验室);Stony Brook University(石溪大学);National University of Singapore(新加坡国立大学)
Multinex: Lightweight Low-light Image Enhancement via Multi-prior Retinex
Department of Computer Science, University of Manchester(曼彻斯特大学计算机科学系);ETcTI, University Politehnica Timisoara(蒂米șoara工业大学ETcTI);West University of Timisoara(蒂米șoara西大学)
DA-DRN: Degradation-Aware Deep Retinex Network for Low-Light Image Enhancement
MOE Key Lab for Neuroinformation, University of Electronic Science and Technology of China (UESTC)(电子科技大学神经信息教育部重点实验室(UESTC));University of Electronic Science and Technology of China (UESTC)(电子科技大学(UESTC))
在真实世界的低光条件下获得的图像不仅亮度低,而且还遭受许多其他类型的退化,例如颜色失真,未知噪声,细节丢失和光晕伪影。在本文中,我们提出了一种用于弱光图像增强的Degradation-Aware Deep Retinex Network(DA-DRN),并解决了上述退化问题。基于Retinex理论,该模型中的分解网络可以将低光照图像分解为反射率图和光照图,并直接处理分解阶段反射率的退化。我们提出了一个退化感知模块(DA模块),它可以指导分解器的训练过程,并使分解器在训练阶段是一个恢复器,而无需在测试阶段增加额外的计算成本。DA模块可以实现去除噪声的目的,同时将细节信息保留到光照图中,并解决颜色失真和光晕伪影。我们引入感知损失来训练增强网络,以生成更符合人类视觉感知的亮度改进的照明图。我们训练和评估我们提出的模型在LOL真实世界和LOL合成数据集上的性能,我们还测试了我们的模型在其他几个常用的数据集没有地面真理(LIME,DICM,MEF和NPE数据集)。我们进行了大量的实验,以证明我们的方法取得了良好的鲁棒性和泛化效果,并优于许多其他国家的最先进的方法定性和定量。我们的方法只需要7毫秒处理一个图像与600 x400分辨率上的TITAN Xp GPU。
ERetinex: Event Camera Meets Retinex Theory for Low-Light Image Enhancement
School of Software Engineering, Xi’an Jiaotong University(西安交通大学软件学院);College of Computer Science and Technology, Jilin University(吉林大学计算机科学与技术学院);Institute of Artificial Intelligence and Robotics, Xi’an Jiaotong University(西安交通大学人工智能与机器人研究所);School of Software, Tsinghua University(清华大学软件学院)
RSEND: Retinex-based Squeeze and Excitation Network with Dark Region Detection for Efficient Low Light Image Enhancement
Department of Computer Science and Engineering University of California, San Diego(计算机科学与工程系,加州大学圣地亚哥分校);Department of Electrical Engineering and Computer Science University of California, Irvine(电气工程与计算机科学系,加州大学伊尔市分校)
A Frequency-Aware Self-Supervised Learning for Ultra-Wide-Field Image Enhancement
College of Information Engineering, Zhejiang University of Technology, Hangzhou, China(浙江工业大学信息工程学院);Ningbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo, China(中国科学院宁波材料技术与工程研究所);Department of Eye and Vision Science, University of Liverpool, Liverpool, UK(利物浦大学眼科学与视觉科学系);School of Biomedical Engineering, ShanghaiTech University, Shanghai, China(上海科技大学生物医学工程学院)
Better Than Reference In Low Light Image Enhancement: Conditional Re-Enhancement Networks
Harbin Institute of Technology(哈尔滨工业大学)
弱光图像存在噪声严重、亮度低、对比度低等问题。在以往的研究中,提出了许多图像增强方法,但很少有方法能同时处理这些问题。针对这些问题,本文提出了一种结合监督学习和以往基于HSV(Hue,Saturation,Value)或Retinex模型的图像增强方法的弱光图像增强方法。首先,分析了HSV颜色空间与Retinex理论的关系,证明了增强图像的V通道(HSV颜色空间中的V通道,等于RGB颜色空间中的最大通道)能够很好地表征对比度和亮度增强过程。然后,提出了一种数据驱动的条件再增强网络(表示为CRENet)。该网络以弱光图像为输入,以增强后的V通道为条件,重新增强弱光图像的对比度和亮度,同时降低噪声和色彩失真。需要注意的是,在训练过程中,任何曝光时间不同的配对图像都可以用于训练,不需要仔细选择监督图像,这会节省很多。此外,在2080 Ti GPU上处理分辨率为400*600的彩色图像所需时间不到20 ms。最后通过对比实验验证了该方法的有效性。实验结果表明,本文提出的方法能够显著提高增强后图像的质量,并且通过与其他图像对比度增强方法相结合,最终增强结果甚至可以在对比度和亮度上优于参考图像。(Code将在https://github.com/hitzhangyu/image-enhancement-with-denoise上提供)