A Bio-Inspired Multi-Exposure Fusion Framework for Low-light Image Enhancement

By Zhenqiang Ying, Ge Li, Wen Gao. arXiv.org, 2017

298 citations recorded by the source corpus.

Low-light images are not conducive to human observation and computer vision algorithms due to their low visibility. Although many image enhancement techniques have been proposed to solve this problem, existing methods inevitably introduce contrast under- and over-enhancement. Inspired by human visual system, we design a multi-exposure fusion framework for low-light image enhancement. Based on the framework, we propose a dual-exposure fusion algorithm to provide an accurate contrast and lightness enhancement. Specifically, we first design the weight matrix for image fusion using illumination estimation techniques. Then we introduce our camera response model to synthesize multi-exposure images. Next, we find the best exposure ratio so that the synthetic image is well-exposed in the regions where the original image is under-exposed. Finally, the enhanced result is obtained by fusing the input image and the synthetic image according to the weight matrix. Experiments show that our method can obtain results with less contrast and lightness distortion compared to that of several state-of-the-art methods.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/b552e69a9c5016d7dd5b5e003c55af84a383bdaa

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

1. Wenhan Yang
2. Jiaying Liu
3. Chen Change Loy
4. Xiaojie Guo
5. Ming-Hsuan Yang
6. Luc Van Gool
7. Jiaya Jia
8. Jiayi Ma

In Media Technology

1. Jiayi Ma
2. Xiaojie Guo
3. Wenhan Yang
4. Jiaying Liu
5. Wenjing Wang
6. Kaihao Zhang
7. Wenhan Luo
8. B. Stenger

Citations in the corpus

• 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.”
• Benchmarking Low-Light Image Enhancement and Beyond (2021, methodology): “For real images, BPDHE, BIMEF, JED, SRIE, Robust, and DeepUPE’s results are under-exposed.”
• Beyond Brightening Low-light Images (2021, methodology): “The state-of-the-art methods of BIMEF (Ying et al. 2017), SRIE (Fu et al. 2016), CRM (Ying et al. 2018), Dong (Dong et al. 2011), LIME (Guo et al. 2017), MF (Fu et al. 2016), RRM (Li et al. 2018), Retinex-Net (Wei et al. 2018), DUPE (Wang et al. 2019), DPE (Chen et al. 2018), GLAD (Wang et al. 2018)”
• R2RNet: Low-light Image Enhancement via Real-low to Real-normal Network (2021, methodology): “BIMEF [14] used a dual-exposure algorithm for image enhancement.”
• An Experiment-Based Review of Low-Light Image Enhancement Methods (2020, methodology): “Ying et al. [223] proposed a novel bioinspired enhancement model in which the source image is generated on the basis of a simulated CRM, and the weight matrix for image fusion is designed using illumination estimation techniques [224], [225].”
• Ultra-High-Definition Low-Light Image Enhancement: A Benchmark and Transformer-Based Method (2022, methodology): “We measure the impact of the proposed axis-based multi-head self attention and dual gated feed-forward network (FFN), see Table 3 (Ying, Li, and Gao 2017) 13.8752 0.5950 0.3264 0.2063 17.9683 0.7972 0.1398 0.1134 FEA (Dong et al. 2011) 16.7165 0.4784 0.3847 0.1421 15.2342 0.7161 0.1949 0.1512 LIME…”
• Learning Enriched Features for Fast Image Restoration and Enhancement (2022, unknown): “[126] [127] [128] [129] [130] [131] [130] [85] [81] [132] [133] [4] [134] (Ours)”
• Low-Light Image Enhancement with Normalizing Flow (2021, unknown): “(Ying, Li, and Gao 2017) 15.95 0.6386 0.4573 DeepUPE (Wang et al. 2019a) 13.19 0.4902 0.4634 JED (Ren et al. 2018) 16.73 0.6817 0.3899 LIME (Guo, Li, and Ling 2016) 14.07 0.5274 0.4021 SICE (Cai, Gu, and Zhang 2018) 18.06 0.7094 0.5078 LLNet (Lore et al. 2017) 17.57 0.7388 0.4021 SRIE (Fu et al.…”
• Learning Enriched Features for Real Image Restoration and Enhancement (2020, unknown): “Input image LIME [39] CRM [115] Retinex-Net [106] SRIE [33] KinD [122] MIRNet (Ours) Ground-truth Fig.”
• MBLLEN: Low-Light Image/Video Enhancement Using CNNs (2018, methodology): “Input MF [11] LIME [14] Ying [43] BIMEF [41] Ours In Table 3, our method almost achieves all the best results under all quality metrics.”
• GLADNet: Low-Light Enhancement Network with Global Awareness (2018, methodology): “Fusion based methods [7] introduce the fusion mechanism of human visual system to help build an accurate image enhancement algorithm.”
• Attention Guided Low-Light Image Enhancement with a Large Scale Low-Light Simulation Dataset (2019, methodology): “BIMEF [82] proposes a dual-exposure fusion algorithm and Ying et al.”
• Kindling the Darkness: A Practical Low-light Image Enhancer (2019, methodology): “The state-of-the-art methods of BIMEF [32], SRIE [12], CRM [33], Dong [34], LIME [11], MF [35], RRM [13], Retinex-Net [21], GLAD [36], MSR [8] and NPE [9] are involved as the competitors.”
• LR3M: Robust Low-Light Enhancement via Low-Rank Regularized Retinex Model (2020, methodology): “…histogram equalization (HE), contrast-limited adaptive histogram equalization (CLAHE) [3], bio-inspired multi-exposure fusion framework (BIMEF) [24] and state-of-the-art Retinex based enhancement methods, i.e. , simultaneous reflectance and illumination estimation (SRIE) [15], naturalness…”