Enhanced Pix2pix Dehazing Network

By Yanyun Qu, Yizi Chen, Jingying Huang, Yuan Xie. Computer Vision and Pattern Recognition, 2019

696 citations recorded by the source corpus.

In this paper, we reduce the image dehazing problem to an image-to-image translation problem, and propose Enhanced Pix2pix Dehazing Network (EPDN), which generates a haze-free image without relying on the physical scattering model. EPDN is embedded by a generative adversarial network, which is followed by a well-designed enhancer. Inspired by visual perception global-first theory, the discriminator guides the generator to create a pseudo realistic image on a coarse scale, while the enhancer following the generator is required to produce a realistic dehazing image on the fine scale. The enhancer contains two enhancing blocks based on the receptive field model, which reinforces the dehazing effect in both color and details. The embedded GAN is jointly trained with the enhancer. Extensive experiment results on synthetic datasets and real-world datasets show that the proposed EPDN is superior to the state-of-the-art methods in terms of PSNR, SSIM, PI, and subjective visual effect.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/0089d5b7e386ae6fccef17fb3b7df5c33b9588a4

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. Ming-Hsuan Yang
2. Luc Van Gool
3. Chen Change Loy
4. Radu Timofte
5. Chongyi Li
6. Chunle Guo
7. Yu Qiao
8. Lei Zhang

Citations in the corpus

• FFA-Net: Feature Fusion Attention Network for Single Image Dehazing (2019, background): “EPDN (Qu et al. 2019) is embedded by a generative adversarial network, which is followed by a welldesigned enhancer without relying on the physical scattering model.”
• Contrastive Learning for Compact Single Image Dehazing (2021, methodology): “Previous works [35, 30, 34, 23, 10, 42] usually employ the fixed grid kernel (e.”
• Domain Adaptation for Image Dehazing (2020, methodology): “The dehazing results by EPDN [25] are also darker than the ground-truths in some cases, as shown in Figure 5 (g).”
• FD-GAN: Generative Adversarial Networks with Fusion-discriminator for Single Image Dehazing (2020, background): “2018) and EPDN (Qu et al. 2019) remove the haze well but bring about exorbitant contrast.”
• Learning to Enhance Low-Light Image via Zero-Reference Deep Curve Estimation (2021, background): “It has also been used to assess the performance of other image restoration tasks, such as image dehazing [47].”
• You Only Look Yourself: Unsupervised and Untrained Single Image Dehazing Neural Network (2020, methodology): “Evaluation Metrics: Like (Irani, 2019; Li et al., 2017; Qu et al., 2019; Ren et al., 2016; Zhang and Patel, 2018), two popular metrics are used in the quantitative comparisons, i.”
• Defocus Blur Detection via Depth Distillation (2020, methodology): “For example, the widely used (atrous) spatial pooling pyramid (SPP [42] or ASPP [2]) has been successful in semantic segmentation and other related tasks [41,31].”
• Zero-Reference Deep Curve Estimation for Low-Light Image Enhancement (2020, background): “It has also been used to assess the performance of other image restoration tasks, such as image dehazing [22].”
• PSD: Principled Synthetic-to-Real Dehazing Guided by Physical Priors (2021, methodology): “(g) EPDN [12] (h) DAD [13] (i) PSD Figure 11: Visual comparison on a real hazy image released from authors of previous work [4–6].”
• All-In-One Image Restoration for Unknown Corruption (2022, background): “Other image restoration tasks have also faced the similar challenge, such as debluring [2,12,29,32–34,36], deraining [10,17,24,42,46,49, 50,52], and dehazing [1,15,20,25,28,35,37,38].”
• Multi-Scale Boosted Dehazing Network With Dense Feature Fusion (2020, methodology): “albedo ambiguity [36], they usually lead to results with significant color distortions when the estimated atmospheric lights or transmission maps are not accurate enough. On the other hand, end-to-end [35, 50, 61, 43, 9, 41, 40, 47, 13] dehazing networks have been proposed to directly restore clean r”