Multi-scale adversarial network for underwater image restoration

Lu, Jingyu; Li, Na; Zhang, Shaoyong; Yu, Zhibin; Zheng, Haiyong; Zheng, Bing

doi:10.1016/j.optlastec.2018.05.048

Cited by 120 publications

(61 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our experiments, the baseline CycleGAN's outputs look somewhat worse than samples presented in the previous works [17,18]. This may be because of the small-quantity and biased training dataset.…”

Section: Discussionmentioning

confidence: 54%

“…By constructing unpaired dataset using turbid underwater images and clean in-air images, Li et al [17] designed an underwater image enhancement network without leveraging the underwater image formation model. Lu et al [18] found direct mapping between turbid and clean underwater images via CycleGAN while estimating the medium transmission via the DCP to improve underwater image quality. Fabbri et al [19] just exploited the CycleGAN to transfer clean underwater images to turbid underwater images then these synthesized pair of images are used for supervised training of conditional GAN.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Adaptive Weighted Multi-Discriminator CycleGAN for Underwater Image Enhancement

Park

Han

2019

JMSE

View full text Add to dashboard Cite

In this paper, we propose a novel underwater image enhancement method. Typical deep learning models for underwater image enhancement are trained by paired synthetic dataset. Therefore, these models are mostly effective for synthetic image enhancement but less so for real-world images. In contrast, cycle-consistent generative adversarial networks (CycleGAN) can be trained with unpaired dataset. However, performance of the CycleGAN is highly dependent upon the dataset, thus it may generate unrealistic images with less content information than original images. A novel solution we propose here is by starting with a CycleGAN, we add a pair of discriminators to preserve contents of input image while enhancing the image. As a part of the solution, we introduce an adaptive weighting method for limiting losses of the two types of discriminators to balance their influence and stabilize the training procedure. Extensive experiments demonstrate that the proposed method significantly outperforms the state-of-the-art methods on real-world underwater images.

show abstract

Section: Discussionmentioning

confidence: 54%

Section: Introductionmentioning

confidence: 99%

Adaptive Weighted Multi-Discriminator CycleGAN for Underwater Image Enhancement

Park

Han

2019

JMSE

View full text Add to dashboard Cite

show abstract

“…However, certain amount of turbidity and color distortion were not removed. Jingyu Lu et al who designed a Dark Channel Prior (DCP) algorithm not only to ensure flexibility but also to improve image restoration performance [6]. Due to occurrence of optical absorption and scattering, Target detection of underwater images is considered to be the most demanding issue.…”

Section: Related Workmentioning

confidence: 99%

Retinex Color Balanced Piecewise Contrast and Fuzzy Trilateral Filter for Underwater Image Enhancement

Parameswari*,

Srinath

2019

IJRTE

View full text Add to dashboard Cite

 Abstract: Over the past few years, underwater observation has become an active research area. Due to the higher rate of image degradation in the underwater environment, image enhancement has become one of the problems to be addressed for the underwater research. Underwater images face limitations like color correction, white balance, color contrast and haze. To overcome those problems, a novel fusion method based on the Retinex Color-balanced Piecewise-contrast and Fuzzy Reinforced Trilateral Filter (RCP-FRTF) method is presented for underwater image improvement. With the underwater image given as input, to start with, a color correction model based on the Retinex multi proportions is presented. With the color corrected output obtained, an Eigen-based White Balancing method is applied to generate color balanced model. With the color balanced underwater image, color contrasting is performed using the Piecewise Linear Color Contrast model. After obtaining the latter, the contrast is said to be improved to a better level. Finally, to generate a haze-free image a Fuzzy Reinforced Trilateral filter is applied. The enhanced and de-hazed images are distinguished by reduced noise level, thus enhanced visibility and contrast while the finest edges are enhanced. The proposed RCP-FRTF method provides better performance in terms of PSNR, computational time, complexity and accuracy as compared to conventional methods.

show abstract

“…Experiments on real-life data show that this method outperforms competing solutions based on the DCP. Another method that relies in part on the DCP method is presented in [26], where an underwater image restoration method is presented based on transferring an underwater style image into a recovered style using Multi-Scale Cycle Generative Adversarial Network System. There, a Structural Similarity Index Measure loss is used to improve underwater image quality.…”

Section: Introductionmentioning

confidence: 99%

Guidelines for Underwater Image Enhancement Based on Benchmarking of Different Methods

Mangeruga

Bruno

Cozza³

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

Images obtained in an underwater environment are often affected by colour casting and suffer from poor visibility and lack of contrast. In the literature, there are many enhancement algorithms that improve different aspects of the underwater imagery. Each paper, when presenting a new algorithm or method, usually compares the proposed technique with some alternatives present in the current state of the art. There are no studies on the reliability of benchmarking methods, as the comparisons are based on various subjective and objective metrics. This paper would pave the way towards the definition of an effective methodology for the performance evaluation of the underwater image enhancement techniques. Moreover, this work could orientate the underwater community towards choosing which method can lead to the best results for a given task in different underwater conditions. In particular, we selected five well-known methods from the state of the art and used them to enhance a dataset of images produced in various underwater sites with different conditions of depth, turbidity, and lighting. These enhanced images were evaluated by means of three different approaches: objective metrics often adopted in the related literature, a panel of experts in the underwater field, and an evaluation based on the results of 3D reconstructions.

show abstract

Multi-scale adversarial network for underwater image restoration

Cited by 120 publications

References 21 publications

Adaptive Weighted Multi-Discriminator CycleGAN for Underwater Image Enhancement

Adaptive Weighted Multi-Discriminator CycleGAN for Underwater Image Enhancement

Retinex Color Balanced Piecewise Contrast and Fuzzy Trilateral Filter for Underwater Image Enhancement

Guidelines for Underwater Image Enhancement Based on Benchmarking of Different Methods

Contact Info

Product

Resources

About