Structure-Revealing Low-Light Image Enhancement Via Robust Retinex Model

By Mading Li, Jiaying Liu, Wenhan Yang, Xiaoyan Sun, Zongming Guo. IEEE Transactions on Image Processing, 2018

1,113 citations recorded by the source corpus.

Low-light image enhancement methods based on classic Retinex model attempt to manipulate the estimated illumination and to project it back to the corresponding reflectance. However, the model does not consider the noise, which inevitably exists in images captured in low-light conditions. In this paper, we propose the robust Retinex model, which additionally considers a noise map compared with the conventional Retinex model, to improve the performance of enhancing low-light images accompanied by intensive noise. Based on the robust Retinex model, we present an optimization function that includes novel regularization terms for the illumination and reflectance. Specifically, we use $\ell _{1}$ norm to constrain the piece-wise smoothness of the illumination, adopt a fidelity term for gradients of the reflectance to reveal the structure details in low-light images, and make the first attempt to estimate a noise map out of the robust Retinex model. To effectively solve the optimization problem, we provide an augmented Lagrange multiplier based alternating direction minimization algorithm without logarithmic transformation. Experimental results demonstrate the effectiveness of the proposed method in low-light image enhancement. In addition, the proposed method can be generalized to handle a series of similar problems, such as the image enhancement for underwater or remote sensing and in hazy or dusty conditions.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/ead9732d17af17a89243f840643563b3bdf14525

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Advancing Image Understanding in Poor Visibility Environments: A Collective Benchmark Study (2019, methodology): “[20] further extended the traditional Retinex model to a robust one with an explicit noise term, and made the first attempt to estimate a noise map out of that model via an alternating direction minimization algorithm.”
LR3M: Robust Low-Light Enhancement via Low-Rank Regularized Retinex Model (2020, methodology): “This dataset contains images provided by authors of [16], [19] and some”
Internet of Underwater Things and Big Marine Data Analytics—A Comprehensive Survey (2020, background): “tortion) and noise imposed by both natural and artificial light sources [193].”
Empowering Things With Intelligence: A Survey of the Progress, Challenges, and Opportunities in Artificial Intelligence of Things (2020, methodology): “Since low-light images usually contain noises that will be ampliﬁed after enhancement, some robust Retinex models have been proposed to account for noise and estimate reﬂectance, illumination, and noise simultaneously [147], [148].”
Learning to Enhance Low-Light Image via Zero-Reference Deep Curve Estimation (2021, methodology): “[20] (d) LightenNet [39] (e) MBLLEN [23] lightenNet MBLLEN Li leida Wang CVPR19 Zero-DCE++”
Underwater Image Enhancement With Hyper-Laplacian Reflectance Priors (2022, positive): “Li et al. [14] made an attempt to estimate the noise map in a robust retinex model, which adopted a ﬁdelity term for the ﬁrst-order gradient of the reﬂectance.”
URetinex-Net: Retinex-based Deep Unfolding Network for Low-light Image Enhancement (2022, methodology): “For LLIE, Gamma correction is widely used to brighten up the illumination map [9, 22, 30], i.e., ˜ L = L 1 , where ˜ L represents adjusted map and the changeable factor � is empirically set to 1 / 2 .”
Retinex-inspired Unrolling with Cooperative Prior Architecture Search for Low-light Image Enhancement (2020, positive): “There are also various classical methods consider Retinex theory [26] and introduce different prior regularized optimization models to characterize the structures of the illumination and reflectance image layers [6, 9, 13, 46, 20].”
SNR-Aware Low-light Image Enhancement (2022, methodology): “Methods Dong [9] LIME [15] MF [11] SRIE [12] BIMEF [55] DRD [45] RRM [28] SID [5] DeepUPE [40] KIND [61] DeepLPF [33] FIDE [48] PSNR 17.”
Beyond Brightening Low-light Images (2021, background): “Following (Guo et al. 2017), Li et al. (2018) further introduced an extra term to host noise.”
Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement (2020, methodology): “[15] proposed a new Retinex model that takes noise into consideration.”
From Fidelity to Perceptual Quality: A Semi-Supervised Approach for Low-Light Image Enhancement (2020, methodology): “Many works [7, 9, 8, 18, 25] impose priors on the decomposed illumination and reflectance.”
STAR: A Structure and Texture Aware Retinex Model (2019, background): “Variational Methods [4]–[11], [20]–[31] have been proposed for Retinex based illumination and reflectance estimation.”
Learning a Simple Low-Light Image Enhancer from Paired Low-Light Instances (2023, methodology): “[17] improved the performance of LIE by introducing a noise map in the Retinex model.”
EnlightenGAN: Deep Light Enhancement Without Paired Supervision (2019, background): “[23] further proposed a robust Retinex model, which additionally considered a noise map compared with the conventional Retinex model, to improve the performance of enhancing low-light images accompanied by intensive noise.”
Sparse Gradient Regularized Deep Retinex Network for Robust Low-Light Image Enhancement (2021, methodology): “Besides RRM [10], JED [11] and RetinexNet, other methods are processed by low-light enhancement methods in the framework proposed by [59]. especially in regions with higher illumination.”