Low-Light Image and Video Enhancement Using Deep Learning: A Survey
This repository provides:
1) a unified online platform, LLIE-Platform http://mc.nankai.edu.cn/ll/, that covers many popular deep learning-based LLIE methods, of which the results can be produced through a user-friendly web interface, contains a low-light image and video dataset.
2) a new dataset LLIV-Phone https://drive.google.com/file/d/1QS4FgT5aTQNYy-eHZ_A89rLoZgx_iysR/view?usp=sharing, in which the images and videos are taken by various phones' cameras under diverse illumination conditions and scenes, and
3) collects deep learning-based low-light image and video enhancement methods, datasets, and evaluation metrics.
More content and details can be found in our Survey Paper: Low-Light Image and Video Enhancement Using Deep Learning: A Survey.
我们提供中文翻译版:基于深度学习的低照度图像与视频增强综述.
We provide the comparison results on the real low-light videos taken by different mobile phones’ cameras at YouTube https://www.youtube.com/watch?v=Elo9TkrG5Oo&t=6s.
✌We will periodically update the content. Welcome to let us know if we miss your work that is published in top-tier Journal or conference. We will add it.
✌Our LLIE-Platform supports the function of download. Please right click and then save the figure.
✌ If you use this dataset or platform, please cite our paper. Please hit the star at the top-right corner. Thanks!
✌ Please hit the star in the repo when you ask for any code. Thanks!
📣News
-
The survey is accepted by TPAMI.
-
We newly add the Zero-DCE++ to the LLIE-Platform. Have Fun!
Zero-DCE++: C. Li, C. Guo, and C. C. Loy, Learning to enhance low-light image via zero-reference deep curve estimation, TPAMI, 2021.
🌱Contents
📋LoLi-Platform
Currently, the LLIE-Platform covers 14 popular deep learning-based LLIE methods including LLNet, LightenNet, Retinex-Net, EnlightenGAN, MBLLEN, KinD, KinD++, TBEFN, DSLR, DRBN, ExCNet, Zero-DCE, Zero-DCE++, and RRDNet, where the results of any inputs can be produced through a user-friendly web interface. Have fun: LLIE-Platform.
📋LLIV-Phone
LLIV-Phone dataset contains 120 videos (45,148 images) taken by 18 different phones' cameras including iPhone 6s, iPhone 7, iPhone7 Plus, iPhone8 Plus, iPhone 11, iPhone 11 Pro, iPhone XS, iPhone XR, iPhone SE, Xiaomi Mi 9, Xiaomi Mi Mix 3, Pixel 3, Pixel 4, Oppo R17, Vivo Nex, LG M322, OnePlus 5T, Huawei Mate 20 Pro under diverse illumination conditions (e.g., weak illumination, underexposure, dark, extremely dark, back-lit, non-uniform light, color light sources, etc.) in the indoor and outdoor scenes.
Anyone can access the LLIV-Phone dataset via
Google Drive: https://drive.google.com/file/d/1QS4FgT5aTQNYy-eHZ_A89rLoZgx_iysR/view?usp=sharing or
Baidu Cloud:https://pan.baidu.com/s/1-8PF3dfbtlHlmk9y5ZKx_w, Password: s0b9)
📋Methods
Date | Publication | Title | Abbreviation | Code | Platform |
---|---|---|---|---|---|
2017 | PR | LLNet: A deep autoencoder approach to natural low-light image enhancement paper | LLNet | Code | Theano |
2018 | PRL | LightenNet: A convolutional neural network for weakly illuminated image enhancement paper | LightenNet | Code | Caffe & MATLAB |
2018 | BMVC | Deep retinex decomposition for low-light enhancement paper | Retinex-Net | Code | TensorFlow |
2018 | BMVC | MBLLEN: Low-light image/video enhancement using CNNs paper | MBLLEN | Code | TensorFlow |
2018 | TIP | Learning a deep single image contrast enhancer from multi-exposure images paper | SCIE | Code | Caffe & MATLAB |
2018 | CVPR | Learning to see in the dark paper | Chen et al. | Code | TensorFlow |
2018 | NeurIPS | DeepExposure: Learning to expose photos with asynchronously reinforced adversarial learning paper | DeepExposure | TensorFlow | |
2019 | ICCV | Seeing motion in the dark paper | Chen et al. | Code | TensorFlow |
2019 | ICCV | Learning to see moving object in the dark paper | Jiang and Zheng | Code | TensorFlow |
2019 | CVPR | Underexposed photo enhancement using deep illumination estimation paper | DeepUPE | Code | TensorFlow |
2019 | ACMMM | Kindling the darkness: A practical low-light image enhancer paper | KinD | Code | TensorFlow |
2019 | ACMMM (IJCV) | Kindling the darkness: A practical low-light image enhancer paper (Beyond brightening low-light images paper) | KinD (KinD++) | Code | TensorFlow |
2019 | ACMMM | Progressive retinex: Mutually reinforced illumination-noise perception network for low-light image enhancement paper | Wang et al. | Caffe | |
2019 | TIP | Low-light image enhancement via a deep hybrid network paper | Ren et al. | Caffe | |
2019(2021) | arXiv(TIP) | EnlightenGAN: Deep light enhancement without paired supervision paper arxiv | EnlightenGAN | Code | PyTorch |
2019 | ACMMM | Zero-shot restoration of back-lit images using deep internal learning paper | ExCNet | Code | PyTorch |
2020 | CVPR | Zero-reference deep curve estimation for low-light image enhancement paper | Zero-DCE | Code | PyTorch |
2020 | CVPR | From fidelity to perceptual quality: A semi-supervised approach for low-light image enhancement paper | DRBN | Code | PyTorch |
2020 | ACMMM | Fast enhancement for non-uniform illumination images using light-weight CNNs paper | Lv et al. | TensorFlow | |
2020 | ACMMM | Integrating semantic segmentation and retinex model for low light image enhancement paper | Fan et al. | ||
2020 | CVPR | Learning to restore low-light images via decomposition-and-enhancement paper | Xu et al. | Code | PyTorch |
2020 | AAAI | EEMEFN: Low-light image enhancement via edge-enhanced multi-exposure fusion network paper | EEMEFN | PyTorch | |
2020 | TIP | Lightening network for low-light image enhancement paper | DLN | PyTorch | |
2020 | TMM | Luminance-aware pyramid network for low-light image enhancement paper | LPNet | PyTorch | |
2020 | ECCV | Low light video enhancement using synthetic data produced with an intermediate domain mapping paper | SIDGAN | TensorFlow | |
2020 | TMM | TBEFN: A two-branch exposure-fusion network for low-light image enhancement paper | TBEFN | Code | TensorFlow |
2020 | ICME | Zero-shot restoration of underexposed images via robust retinex decomposition paper | RRDNet | Code | PyTorch |
2020 | TMM | DSLR: Deep stacked laplacian restorer for low-light image enhancement paper | DSLR | Code | PyTorch |
2021 | TPAMI | Learning to enhance low-light image via zero-reference deep curve estimation paper | Zero-DCE++ | Code | PyTorch |
2021 | CVPR | Retinex-inspired unrolling with cooperative prior architecture search for low-light image enhancement paper | RUAS | Code | PyTorch |
2021 | CVPR | Learning temporal consistency for low light video enhancement from single images paper | Zhang et al. | Code | PyTorch |
2021 | CVPR | Nighttime visibility enhancement by increasing the dynamic range and suppression of light effects paper | Sharma and Tan | ||
2021 | TCSVT | RetinexDIP: A unified deep framework for low-light image enhancement paper | RetinexDIP | Code | PyTorch |
2021 | TIP | Sparse gradient regularized deep retinex network for robust low-light image enhancement paper | Retinex-Net | PyTorch | |
2021 | TCSVT | Low-light image enhancement via progressive-recursive network paper | PRIEN | PyTorch | |
2021 | TIP | Band representation-based semi-supervised low-light image enhancement: Bridging the gap between signal fidelity and perceptual quality paper | DRBN | PyTorch | |
2021 | ICCV | Adaptive unfolding total variation network for low-light image enhancement paper | UTVNet | Code | PyTorch |
2021 | ICCV | Seeing dynamic scene in the dark: A high-quality video dataset with mechatronic alignment paper | Wang et al. | Code | PyTorch |
2022 | CVPR | SNR-aware Low-Light Image Enhancement paper | SNR-Aware | Code | PyTorch |
2022 | CVPR | Toward Fast, Flexible, and Robust Low-Light Image Enhancement paper | SCI | Code | PyTorch |
2022 | CVPR | URetinex-Net: Retinex-Based Deep Unfolding Network for Low-Light Image Enhancement paper | URetinex-Net | Code | PyTorch |
📋Datasets
Abbreviation | Number | Format | Real/Synetic | Video | Paired/Unpaired/Application | Dataset |
---|---|---|---|---|---|---|
LOL paper | 500 | RGB | Real | No | Paired | Dataset |
SCIE paper | 4413 | RGB | Real | No | Paired | Dataset |
MIT-Adobe FiveK paper | 5000 | Raw | Real | No | Paired | Dataset |
SID paper | 5094 | Raw | Real | No | Paired | Dataset |
DRV paper | 202 | Raw | Real | Yes | Paired | Dataset |
SMOID paper | 179 | Raw | Real | Yes | Paired | Dataset |
SDSD paper | 150 | RGB | Real | Yes | Paired | Dataset |
LIME paper | 10 | RGB | Real | No | Unpaired | Dataset |
NPE paper | 84 | RGB | Real | No | Unpaired | Dataset |
MEF paper | 17 | RGB | Real | No | Unpaired | Dataset |
DICM paper | 64 | RGB | Real | No | Unpaired | Dataset |
VV | 24 | RGB | Real | No | Unpaired | Dataset |
ExDARK paper | 7363 | RGB | Real | No | Application (Object Detection) | Dataset |
BBD-100K paper | 10,000 | RGB | Real | Yes | Application (Driving with diverse kinds of annotations) | Dataset |
DARK FACE paper | 6000 | RGB | Real | No | Application (Face Recognition) | Dataset |
NightCity paper | 4297 | RGB | Real | No | Application (Semantic Segmentation) | Dataset |
📋Metrics
Abbreviation | Full-/Non-Reference | Platform | Code |
---|---|---|---|
MAE (Mean Absolute Error) | Full-Reference | ||
MSE (Mean Square Error) | Full-Reference | ||
PSNR (Peak Signal-to-Noise Ratio) | Full-Reference | ||
SSIM (Structural Similarity Index Measurement) | Full-Reference | MATLAB | Code |
LPIPS (Learned Perceptual Image Patch Similarity) | Full-Reference | PyTorch | Code |
LOE (Lightness Order Error) | Non-Reference | MATLAB | Code |
NIQE (Naturalness Image Quality Evaluator) | Non-Reference | MATLAB | Code |
PI (Perceptual Index) | Non-Reference | MATLAB | Code |
SPAQ (Smartphone Photography Attribute and Quality) | Non-Reference | PyTorch | Code |
NIMA (Neural Image Assessment) | Non-Reference | PyTorch/TensorFlow | Code/Code |
MUSIQ (Multi-scale Image Quality Transformer) | Non-Reference | TensorFlow | Code |
📜License
The code, platform, and dataset are made available for academic research purpose only.
📚Citation
If you find the repository helpful in your resarch, please cite the following paper.
@article{LoLi,
title={Low-Light Image and Video Enhancement Using Deep Learning: A Survey},
author={Li, Chongyi and Guo, Chunle and Han, Linghao and Jiang, Jun and Cheng, Ming-Ming and Gu, Jinwei and Loy, Chen Change},
journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
year={2021}
}