Variational Regularized Transmission Refinement for Image Dehazing

Shu, Qiaoling; Wu, Chuansheng; Xiao, Zhe; Liu, Ryan Wen

doi:10.1109/icip.2019.8803256

Cited by 28 publications

(23 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead, the method of this paper uses the atmospheric multiple scattering physical model, which can avoid the blurring of images caused by multiple scattering, especially in the ocean scene. Second, the method of this paper uses a new network structure, which uses the latest smooth dilation [27] and sub-pixel [26] techniques to avoid gridding artifacts and the halo artifacts, and uses multi-scale sub-network to fuse multiscale feature information, while the AOD-Net method only uses a spanning-connected convolutional neural network as the network model. Third, the loss function is different, we proposed and used multiple loss functions to optimize the network model, specifically, the method of this paper use Mean Square Error loss, multi-scale structural similarity loss [28], and perceptual loss [29], which can not only help the network focus on image details, but also consider the texture and structural information of the image during the training process.…”

Section: Learning-based Methodsmentioning

confidence: 99%

“…Specifically, the proposed network model was based on reconstructed atmospheric multiple scattering model, which combined multiple parameters into one parameter and estimated to dehazing image. In the end-to-end network, subpixel convolution [26] was used instead of transposed convolution to avoid halo artifacts, and smooth dilation convolution [27] was used instead of transposed convolution to avoid gridding artifacts. In the ocean scene, due to the complexity of detail, structure and texture information, multiple loss functions were proposed to optimize the network, which contained Mean Square Error loss, multi-scale structural similarity loss [28] and perceptual loss [29].…”

Section: Introductionmentioning

confidence: 99%

“…• In the end-to-end network, subpixel convolution [26] is used instead of transposed convolution to avoid halo artifacts, and smooth dilation convolution [27] is used instead of transposed convolution to avoid gridding artifacts.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

An end-to-end sea fog removal network using multiple scattering model

Huang

Zheng

et al. 2021

PLoS ONE

View full text Add to dashboard Cite

An end-to-end sea fog removal network using multiple scattering model was proposed. In this network, the atmospheric multiple scattering model was re-formulated and used for sea fog removal. Compared with the atmospheric single scattering model, the atmospheric multiple scattering model could more comprehensively consider the effect of multiple scattering, which was important to the dense fog scenes, such as in ocean scene. Therefore, we used the atmospheric multiple scattering model to avoid image blurring. The model can directly generate the dehazing results, and unify the three parameters of the transmission map, the atmospheric light and the blur kernel into one formula. The latest smooth dilation and sub-pixel techniques were used in the network model. The latest techniques can avoid the gridding artifacts and the halo artifacts, the multi-scale sub-network was used to consider the features of multi-scale. In addition, multiple loss functions were used in end-to-end network. In the experimental results, the model was superior to the state-of-the-art models in terms of quantitatively and qualitatively.

show abstract

Section: Learning-based Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

An end-to-end sea fog removal network using multiple scattering model

Huang

Zheng

et al. 2021

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Since the variational model has a good dehazing effect in the sky area and edge area, many variational methods have been proposed to achieve image dehazing. Shu et al [30] introduced an effective two-step transmission map estimation model, which uses a hybrid variational regularization model to refine the transmission and fuses two different transmission maps to obtain the sharp image. In [31], a hybrid regularized variational framework is proposed, which introduces the secondorder total generalized variation regularizer to constrain the estimation of a depth map.…”

Section: A Single Image Dehazing Methodsmentioning

confidence: 99%

A Mixed Transmission Estimation Iterative Method for Single Image Dehazing

et al. 2021

View full text Add to dashboard Cite

Considering the frequent occurrence of hazy weather, single image dehazing has become an important research task. Physical model-based and deep learning-based methods are two competitive methods in single image dehazing, but achieving fidelity and effectively removing haze at the same time in real hazy scenes is still a challenging problem. In this work, we propose a mixed iterative model to restore high-quality clear images by integrating a physical model-based approach and a learning-based approach, which can combine the advantages of the physical model-based method in maintaining natural attributes and the advantages of the learning-based method in completely removing haze. For a hazy image, first, according to the haze density, it is divided into separate regions to calculate the local atmospheric light. Then, we utilize the dark channel prior and DehazeNet to jointly estimate the transmission, which can estimate an accurate recovered image that is more in line with the real scene. Finally, a numerical iterative strategy is employed to further optimize the atmospheric light and transmission. The experiments on both synthetic and natural hazy image datasets indicate that the proposed method can restore natural and clean images and has superior performance than other state-of-the-art algorithms in peak signal-to-noise ratio, structural similarity, information entropy, color natural index, and color histogram. And in the four indicators of the quantitative results of the synthetic hazy test set, we achieve the highest score in three of them and the second highest result in another. In the comparison of the two quantitative indicators of natural hazy images, our average objective results are both the highest.INDEX TERMS Image dehazing, numerical iterative, mixed model, joint transmission.

show abstract

“…The second-order framework can preserving important structures in both depth map and haze-free image. Shu [14] proposed a hybrid variational model with promoted regularization terms to refining transmission map, and then using an alternating direction algorithm to obtained final haze-free image.…”

Section: Introductionmentioning

confidence: 99%

An End-to-End System for Unmanned Aerial Vehicle High-Resolution Remote Sensing Image Haze Removal Algorithm Using Convolution Neural Network

2020

View full text Add to dashboard Cite

An end-to-end image dehazing method based on convolution neural network is presented to solve the problem in which Unmanned Aerial Vehicle (UAV) high-resolution remote sensing images have reduced image sharpness due to haze. First, the original atmospheric scattering model is adapted to get an end-to-end dehazing model. Then, several unknown parameters are unified into one parameter, and the unknown parameter is estimated by using a multiscale convolution neural network. Finally, the parameter estimates are incorporated into the dehazing model to get a haze-free image. For the no reference image dataset, we first train the network using existing datasets, and then the network is trained using a self-built dataset. In this paper, the haze removal effect for different types of unmanned remote sensing images is tested and compared with those of the main dehazing algorithms. The experiments show that the algorithm in this paper has different degrees of improvement regarding its visual effect and objective indicators. INDEX TERMS End to end image dehazing; Deep learning; Multiscale convolutional neural network; Remote sensing image Yufeng Li, PHD, professor and master's supervisor at the School of electronic information engineering, Shenyang University of Aeronautics and Astronautics. He graduated from Northern Traffic University in July 1991, majoring in communication engineering. In March 2002, he graduated from Jilin University, majoring in communication and information system. In

show abstract

Variational Regularized Transmission Refinement for Image Dehazing

Cited by 28 publications

References 27 publications

An end-to-end sea fog removal network using multiple scattering model

An end-to-end sea fog removal network using multiple scattering model

A Mixed Transmission Estimation Iterative Method for Single Image Dehazing

An End-to-End System for Unmanned Aerial Vehicle High-Resolution Remote Sensing Image Haze Removal Algorithm Using Convolution Neural Network

Contact Info

Product

Resources

About