LIME: Low-Light Image Enhancement via Illumination Map Estimation

By Xiaojie Guo, Yu Li, Haibin Ling. 2017

2,702 citations recorded by the source corpus.

When one captures images in low-light conditions, the images often suffer from low visibility. Besides degrading the visual aesthetics of images, this poor quality may also significantly degenerate the performance of many computer vision and multimedia algorithms that are primarily designed for high-quality inputs. In this paper, we propose a simple yet effective low-light image enhancement (LIME) method. More concretely, the illumination of each pixel is first estimated individually by finding the maximum value in R, G, and B channels. Furthermore, we refine the initial illumination map by imposing a structure prior on it, as the final illumination map. Having the well-constructed illumination map, the enhancement can be achieved accordingly. Experiments on a number of challenging low-light images are present to reveal the efficacy of our LIME and show its superiority over several state-of-the-arts in terms of enhancement quality and efficiency.

Subfields: Computer Vision and Pattern Recognition

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

Citations in the corpus

Refining Language Models with Compositional Explanations (2021, methodology): “As a prominent example, feature attribution methods [42, 14, 19] interpret model prediction on an instance by assigning an importance (attribution) score to each input feature (or token, in the context of NLP tasks), which helps uncover an overemphasis or understatement of a specific feature (e.”
Deep Retinex Decomposition for Low-Light Enhancement (2018, methodology): “s is described in detail in the following section. 2.2 Structure-Aware Smoothness Loss One basic assumption for illumination map is the local consistency and the structure- awareness, as mentioned by [9]. In other words, a good solution for illumination map should be smooth in textural details while”
Learning a Simple Low-Light Image Enhancer from Paired Low-Light Instances (2023, background): “To correct contrast, uncover textures, and remove sensor noise, great efforts have been made in developing Low-light Image Enhancement (LIE) algorithms in the past decades [1, 5, 6, 8, 28, 35].”
STAR: A Structure and Texture Aware Retinex Model (2019, methodology): “We compare the proposed STAR model with previous competing low-light image enhancement methods, including HE [47], MSRCR [19], Contextual and Variational Contrast (CVC) [48], Naturalness Preserved Enhancement (NPE) [5], LDR [49], SIRE [32], MF [50], WVM [7], LIME [6], and JieP [8].”
Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement (2020, methodology): “[8] first estimated a coarse illumination map by searching the maximum intensity of each pixel in RGB channels, then refining the coarse illumination map by a structure prior; Li et al.”
Toward Fast, Flexible, and Robust Low-Light Image Enhancement (2022, positive): “Further, Guo et al. [8] adopted the relative total variation [28] as the constraint of the illumination.”
Structure-Revealing Low-Light Image Enhancement Via Robust Retinex Model (2018, methodology): “(a) is the input image; (b)–(f) are enhancement results of HE, LIME [13], NPE [11], PIE [23], and SRIE [14] with a post-processing performed by BM3D with the denoising parameter σ = 30; (g) is the result obtained by the proposed method with model (6).”
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 challenging data with much noise from [42] and [43].”
Learning to Enhance Low-Light Image via Zero-Reference Deep Curve Estimation (2021, methodology): “We compare our method with several state-of-the-art methods: three conventional methods (SRIE [18], LIME [19], Li et al.”
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.”
Advancing Image Understanding in Poor Visibility Environments: A Collective Benchmark Study (2019, background): “We also include seven state-of-the-art algorithms for light/contrast enhancement: (a) Bio-Inspired Multi-Exposure Fusion (BIMEF) [154]14; (b) Dehazing [155](14); (c) Low-light IMage Enhancement (LIME) [108]15; (d) MF [107](14); (e)”
Retinex-inspired Unrolling with Cooperative Prior Architecture Search for Low-light Image Enhancement (2020, methodology): “With the illumination warm-start t̂k, we further refine t by solving the following model inspired by the work in [9]: mint 1 2‖t− t̂k‖ (2)+φ(t),where φ(·) represents a regularization term of t.”
Sparse Gradient Regularized Deep Retinex Network for Robust Low-Light Image Enhancement (2021, methodology): “The BM3D is performed as the post-processing, following the operations in LIME [8].”
MSR-net: Low-light Image Enhancement Using Deep Convolutional Network (2017, methodology): “Meth-ods [16],[8],[11],[10] are implemented using CPU, while our method is tested on both CPU and GPU.”
Low-Light Image Enhancement With Semi-Decoupled Decomposition (2020, methodology): “In this perspective, our decomposition model lies between the coupled models [1]–[4], [6] and the decoupled models [7], [8] that only estimates I .”
Underexposed Photo Enhancement Using Deep Illumination Estimation (2019, methodology): “Similar to that of [11, 14], we also regard the reflectance component Ĩ as a well-exposed image, so in our model, we take Ĩ as the enhancement result and I as the observed underexposed image.”
Retinexformer: One-stage Retinex-based Transformer for Low-light Image Enhancement (2023, methodology): “Similar to [15, 18, 49], we regard R as a well-exposed image.”
URetinex-Net: Retinex-based Deep Unfolding Network for Low-light Image Enhancement (2022, methodology): “4 that some models, e.g ., NPE, SRIE, LIME, Retinex-Net and Zero-DCE introduce and even amplify noise after enhancement, such that they suffer from severe noise distortion and color degradation when brightening dark regions.”
Learning a Deep Single Image Contrast Enhancer from Multi-Exposure Images (2018, methodology): “and image analysis algorithms, but also degrade the visual aesthetics of images [1].”