EnlightenGAN: Deep Light Enhancement Without Paired Supervision

Jiang, Yifan; Gong, Xinyu; Liu, Ding; Cheng, Yu; Chen, Fang; Shen, Xiaohui; Yang, Jianchao; Zhou, Pan; Wang, Zhangyang

doi:10.1109/tip.2021.3051462

Cited by 1,369 publications

(887 citation statements)

References 44 publications

Supporting

Mentioning

884

Contrasting

Unclassified

Order By: Relevance

“…To demonstrate the efficiency of our method, we evaluate it qualitative and quantitative on MIT-Adobe5k and LOL and compare our method with seven state-of-the-art CNN-based methods (RetinexNet [4], KinD [5], EnGAN [25], MIRNet [26], PMANet [27], DeepUPE [17], DRBN [18]). For fair comparison, The results are reproduced by publicly-available models released by the authors.…”

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

“…Where l t is the latency threshold, l r is the latency relax range. We propose the relax range because the networks with a la-Input KinD [5] EnGAN [25] DeepUPE [17] ZeroDCE [20] DRBN [18] MIRNet [26] Ours tency slightly higher than the latency threshold usually contain one or two choices shared with low-latency networks. We believe that these choices should also be well trained.…”

Section: Preliminariesmentioning

confidence: 99%

See 1 more Smart Citation

Hardware-Aware Low-Light Image Enhancement via One-Shot Neural Architecture Search with Shrinkage Sampling

Yao

Liu

Zhang

et al. 2021

2021 IEEE International Conference on Multimedia and Expo (ICME)

View full text Add to dashboard Cite

Low-light image enhancement has traditionally been tackled by training a heuristically designed neural network architecture. Despite the success of these approaches, the heuristic design pattern inherently not only hinders further optimization of network architectures, but also limits the factors that the designer can take into consideration. As a result, these methods are difficult to achieve a balance between enhancing performance and hardware related performance. In this paper, we equip a basic enhancing algorithm with a neural architecture search technique. This technique helps to automatically search an optimal hardware-aware architecture while also increases neglectable computation burden. In this work, we propose a shrinkage sampling strategy to drastically decrease the computation cost of neural architecture search while improving the quality of search. Extensive experiments on various benchmarks demonstrate that our algorithm achieves state-of-the-art performance with higher speed.

show abstract

Section: Comparison With State-of-the-art Methodsmentioning

confidence: 99%

Section: Preliminariesmentioning

confidence: 99%

Hardware-Aware Low-Light Image Enhancement via One-Shot Neural Architecture Search with Shrinkage Sampling

Yao

Liu

Zhang

et al. 2021

2021 IEEE International Conference on Multimedia and Expo (ICME)

View full text Add to dashboard Cite

show abstract

“…Guo et al [2] set a series of nonreference loss functions to enable the network to perform end-to-end training without any reference images. Jiang et al [18] proposed an efficient and unsupervised generative adversarial network called EnlightenGAN for the low-light image enhancement problem, which can be trained without low/normal-light image pairs. N. Anantrasirichai and David Bull [19] used an adaptation of the CycleGan structure to color and denoise images.…”

Section: Related Workmentioning

confidence: 99%

Multi-Feature Guided Low-Light Image Enhancement

Hong

Shao

et al. 2021

Applied Sciences

View full text Add to dashboard Cite

Due to the characteristics of low signal-to-noise ratio and low contrast, low-light images will have problems such as color distortion, low visibility, and accompanying noise, which will cause the accuracy of the target detection problem to drop or even miss the detection target. However, recalibrating the dataset for this type of image will face problems such as increased cost or reduced model robustness. To solve this kind of problem, we propose a low-light image enhancement model based on deep learning. In this paper, the feature extraction is guided by the illumination map and noise map, and then the neural network is trained to predict the local affine model coefficients in the bilateral space. Through these methods, our network can effectively denoise and enhance images. We have conducted extensive experiments on the LOL datasets, and the results show that, compared with traditional image enhancement algorithms, the model is superior to traditional methods in image quality and speed.

show abstract

“…Unsupervised methods. Jiang et al [26] firstly apply unpaired training to the work of underexposed image enhancement, avoiding the dependence on paired data. They construct a dual discriminator structure to process global information and local information separately.…”

Section: Related Workmentioning

confidence: 99%

Underexposed Image Enhancement via Unsupervised Feature Attention Network

2021

2021 IEEE International Conference on Multimedia and Expo (ICME)

View full text Add to dashboard Cite

In order to solve the problem that deep learning method needs a lot of paired data sets in image enhancement, this paper proposes unsupervised feature attention network (UFANet), which uses a new illumination estimation that combines pixel estimation and channel estimation to guide the network to decompose underexposed images. In addition, a feature attention residual network is trained to decompose under-exposed images into illumination and reflectance. Through a set of carefully designed non reference loss functions, which implicitly enhance the quality and drive the learning of the network, we train UFANet without any paired images. A large number of experiments on various benchmarks have proved the advantages of our method over the latest methods in terms of quality and quantity. Compared to the state-of-the-art methods, our method only needs to be trained on 350 underexposed images. All the above advantages make our UFANet attractive in practical applications.

show abstract

EnlightenGAN: Deep Light Enhancement Without Paired Supervision

Cited by 1,369 publications

References 44 publications

Hardware-Aware Low-Light Image Enhancement via One-Shot Neural Architecture Search with Shrinkage Sampling

Hardware-Aware Low-Light Image Enhancement via One-Shot Neural Architecture Search with Shrinkage Sampling

Multi-Feature Guided Low-Light Image Enhancement

Underexposed Image Enhancement via Unsupervised Feature Attention Network

Contact Info

Product

Resources

About