EnlightenGAN: Deep Light Enhancement without Paired Supervision

Abstract

Deep learning-based methods have achieved remarkable success in image restoration and enhancement, but are they still competitive when there is a lack of paired training data? As one such example, this paper explores the low-light image enhancement problem, where in practice it is extremely challenging to simultaneously take a low-light and a normal-light photo of the same visual scene. We propose 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. Instead of supervising the learning using ground truth data, we propose to regularize the unpaired training using the information extracted from the input itself, and benchmark a series of innovations for the low-light image enhancement problem, including a global-local discriminator structure, a self-regularized perceptual loss fusion, and attention mechanism. Through extensive experiments, our proposed approach outperforms recent methods under a variety of metrics in terms of visual quality and subjective user study. Thanks to the great flexibility brought by unpaired training, EnlightenGAN is demonstrated to be easily adaptable to enhancing real-world images from various domains. The code is available at \url{https://github.com/yueruchen/EnlightenGAN}

References in corpus (7)

• Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
• Deep Bilateral Learning for Real-Time Image Enhancement
• The relativistic discriminator: a key element missing from standard GAN
• Deep Retinex Decomposition for Low-Light Enhancement
• PsyPhy: A Psychophysics Driven Evaluation Framework for Visual Recognition
• Improved Techniques for Learning to Dehaze and Beyond: A Collective Study
• UG${}^{2+}$ Track 2: A Collective Benchmark Effort for Evaluating and Advancing Image Understanding in Poor Visibility Environments

Cited by in corpus (20)

• Towards Unsupervised Deep Image Enhancement with Generative Adversarial Network
• Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement
• DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better
• A Switched View of Retinex: Deep Self-Regularized Low-Light Image Enhancement
• Progressive Joint Low-light Enhancement and Noise Removal for Raw Images
• AutoGAN: Neural Architecture Search for Generative Adversarial Networks
• Dehaze-GLCGAN: Unpaired Single Image De-hazing via Adversarial Training
• DarkDeblur: Learning single-shot image deblurring in low-light condition
• Frequency Domain Loss Function for Deep Exposure Correction of Dark Images
• A Retinex based GAN Pipeline to Utilize Paired and Unpaired Datasets for Enhancing Low Light Images
• DALE : Dark Region-Aware Low-light Image Enhancement
• Exploring Image Enhancement for Salient Object Detection in Low Light Images
• ReLLIE: Deep Reinforcement Learning for Customized Low-Light Image Enhancement
• MM-Hand: 3D-Aware Multi-Modal Guided Hand Generative Network for 3D Hand Pose Synthesis
• Deep Relighting Networks for Image Light Source Manipulation
• Intrinsic Image Transfer for Illumination Manipulation
• GIMP-ML: Python Plugins for using Computer Vision Models in GIMP
• Disentangle Perceptual Learning through Online Contrastive Learning
• Shed Various Lights on a Low-Light Image: Multi-Level Enhancement Guided by Arbitrary References
• Unpaired Image Enhancement with Quality-Attention Generative Adversarial Network