Learning Deep Context-Sensitive Decomposition for Low-Light Image Enhancement

@article{Ma2021LearningDC,
  title={Learning Deep Context-Sensitive Decomposition for Low-Light Image Enhancement},
  author={Long Ma and Risheng Liu and Jiaao Zhang and Xin Fan and Zhongxuan Luo},
  journal={IEEE transactions on neural networks and learning systems},
  year={2021},
  volume={PP}
}
Enhancing the quality of low-light (LOL) images plays a very important role in many image processing and multimedia applications. In recent years, a variety of deep learning techniques have been developed to address this challenging task. A typical framework is to simultaneously estimate the illumination and reflectance, but they disregard the scene-level contextual information encapsulated in feature spaces, causing many unfavorable outcomes, e.g., details loss, color unsaturation, and… 
[PDF] Semantic Reader
5 Citations
Highly Influential Citations
1
Background Citations
2
Methods Citations
3

5 Citations

Learning with Nested Scene Modeling and Cooperative Architecture Search for Low-Light Vision
TLDR
Retinex-inspired Unrolling with Architecture Search (RUAS) is established, a general learning framework, which not only can address low-light enhancement task, but also has the flexibility to handle other more challenging downstream vision applications.
Low-light Image Enhancement by Retinex Based Algorithm Unrolling and Adjustment
TLDR
The proposed framework contains a decomposition network inspired by algorithm unrolling, and adjustment networks considering both global brightness and local brightness sensitivity, and a self-supervised fine-tuning strategy, which can always guarantee a promising performance.
Low-light image enhancement based on variational Retinex model
TLDR
A novel variational Retinex model is built to accurately estimate the illumination and reflectance components and outperforms the state-of-the-art methods in RetineX decomposition and illumination adjustment.
Multi-scale low-light image enhancement by fusing dense residual blocks
TLDR
The experimental results show that the proposed method better improves the contrast and brightness, actualizes the color, enriches the texture details, and decreases the noise and artifacts in low-light images, which quantitatively and qualitatively demonstrated its advantages comparing with other mainstream methods.
Target-aware Dual Adversarial Learning and a Multi-scenario Multi-Modality Benchmark to Fuse Infrared and Visible for Object Detection
TLDR
This paper proposes a bilevel optimization formulation for the joint problem of fusion and detection, and then unrolls to a target-aware Dual Adversarial Learning (TarDAL) network for fusion and a commonly used detection network.

References

SHOWING 1-10 OF 60 REFERENCES
Deep Retinex Decomposition for Low-Light Enhancement
TLDR
Extensive experiments demonstrate that the proposed deep Retinex-Net learned on this LOw-Light dataset not only achieves visually pleasing quality for low-light enhancement but also provides a good representation of image decomposition.
Low-Light Image Enhancement via a Deep Hybrid Network
TLDR
A novel spatially variant recurrent neural network (RNN) is proposed as an edge stream to model edge details, with the guidance of another auto-encoder, to enhance the visibility of degraded images.
Beyond Brightening Low-light Images
TLDR
This work builds a simple yet effective network, which, inspired by Retinex theory, decomposes images into two components, followed a divide-and-conquer principle, and is trained with paired images shot under different exposure conditions.
LIME: Low-Light Image Enhancement via Illumination Map Estimation
TLDR
Experiments on a number of challenging low-light images are present to reveal the efficacy of the proposed LIME and show its superiority over several state-of-the-arts in terms of enhancement quality and efficiency.
From Fidelity to Perceptual Quality: A Semi-Supervised Approach for Low-Light Image Enhancement
TLDR
A deep recursive band network (DRBN) is proposed to recover a linear band representation of an enhanced normal-light image with paired low/normal-light images, and then obtain an improved one by recomposing the given bands via another learnable linear transformation based on a perceptual quality-driven adversarial learning with unpaired data.
Learning to Restore Low-Light Images via Decomposition-and-Enhancement
TLDR
A frequency-based decompositionand- enhancement model that first learns to recover image objects in the low-frequency layer and then enhances high-frequency details based on the recovered image objects and outperforms state-of-the-art approaches in enhancing practical noisy low-light images.
Underexposed Photo Enhancement Using Deep Illumination Estimation
TLDR
A new neural network for enhancing underexposed photos is presented, which introduces intermediate illumination in its network to associate the input with expected enhancement result, which augments the network's capability to learn complex photographic adjustment from expert-retouched input/output image pairs.
LLNet: A deep autoencoder approach to natural low-light image enhancement
EnlightenGAN: Deep Light Enhancement Without Paired Supervision
TLDR
This paper proposes a highly effective unsupervised generative adversarial network, dubbed EnlightenGAN, that can be trained without low/normal-light image pairs, yet proves to generalize very well on various real-world test images.
Low-light Image Enhancement Algorithm Based on Retinex and Generative Adversarial Network
TLDR
The proposed GAN-based network and employed Retinex theory have proven to be effective in dealing with the low-light image enhancement problems, which will benefit the image processing with no doubt.
...
...