Underwater image enhancement using improved generative adversarial network

Zhang, Tingting; Li, Yujie; Takahashi, Satoshi

doi:10.1002/cpe.5841

Cited by 10 publications

(11 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The software based approaches helps in recovering underwater images by utilising efficient algorithms. These software approaches can be further divided into image restoration method, colour correction methods, dark channel prior (DCP) based methods, fusion based methods and convolutional neural networks (CNNs) based methods [13][14][15][16][17][18][19][20][21][22][23].…”

Section: Literature Surveymentioning

confidence: 99%

“…Takahashi et al. [22] proposed generative adversarial network for colour restoration of underwater images. The loss function of network has been improved to train the dataset, followed by detecting and enhancing the underwater images.…”

Section: Literature Surveymentioning

confidence: 99%

“…Now a days the generative adversarial networks have also been increasingly used in underwater image enhancement. Takahashi et al [22] proposed generative adversarial network for colour restoration of underwater images. The loss function of network has been improved to train the dataset, followed by detecting and enhancing the underwater images.…”

Section: Literature Surveymentioning

confidence: 99%

See 2 more Smart Citations

Enhancement algorithm for high visibility of underwater images

Mathur

Goel

2021

IET Image Processing

View full text Add to dashboard Cite

Underwater image enhancement has been attracting researchers in present day scenario for exploring the marine life. But underwater images suffer from various glitches like haziness, low contrast and faded colours due to absorption and scattering properties of light in water. To overcome these issues, the present paper proposes an enhancement method for underwater images. The proposed enhancement method comprises of three steps, i.e. automatic white balancing, dehazing and Rayleigh stretching in spatial domain. Automatic white balancing technique removes the colour cast from underwater images. Haze removal algorithm efficiently overcomes the haziness but it also reduces the local contrast and making the image appear dull. For improving the dehazing result and the visual quality simultaneously, the proposed method process the image by histogram stretching technique based on Rayleigh distribution. The proposed method corrects the colour cast and improves the contrast along with removing the haze from underwater images. Subjective and objective analysis on U 45 dataset validates the efficiency of proposed method and the enhanced images achieve better visual quality as compared to the existing methods. High values of EMEE, EME, UIQM and UCIQE for different types of underwater images further proves the potential and efficacy of the proposed method. INTRODUCTIONAn unexplored huge world present under oceans is attracting various researcher and scientists to study and investigate variety of marine life forms, incredible landscape, mysterious shipwrecks and beautiful coral reefs. Underwater environment is also gaining substantial attention due to increase in optical robotic applications and marine engineering [1,2]. But complex underwater environment and poor lighting conditions are major challenges to capture good quality underwater images. Water is a denser medium than the air and sunlight is reflected and refracted partially as it enters the ocean due to transition from one medium to the other. Less light will be reflected when ocean water is calm and smooth rather than when it is turbulent. The part of light which penetrates the ocean surface is affected by the phenomenon of refraction because light travels faster in air as compared to water. These refracted rays of light further interact with the water molecules and suspended solid particles leading to scattering and absorption. Scattering reduces intensity of light causing loss in contrast and haziness in underwater images. Whereas, absorption of This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Section: Literature Surveymentioning

confidence: 99%

Section: Literature Surveymentioning

confidence: 99%

Section: Literature Surveymentioning

confidence: 99%

See 1 more Smart Citation

Enhancement algorithm for high visibility of underwater images

Mathur

Goel

2021

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…However, the detection methods mentioned above can get excellent performance on common vehicle detection, which are not properly suitable for nighttime vehicle detection. To deal with the detection problems in nighttime, GAN [39][40][41] has shown excellent performance on image processing recently. There is a limited number of research that apply the GAN network to nighttime vehicle detection, and some of them have limitations.…”

Section: Related Workmentioning

confidence: 99%

A Nighttime Vehicle Detection Method with Attentive GAN for Accurate Classification and Regression

Liu

Qiu

Wang

et al. 2021

Entropy

View full text Add to dashboard Cite

Vehicle detection plays a vital role in the design of Automatic Driving System (ADS), which has achieved remarkable improvements in recent years. However, vehicle detection in night scenes still has considerable challenges for the reason that the vehicle features are not obvious and are easily affected by complex road lighting or lights from vehicles. In this paper, a high-accuracy vehicle detection algorithm is proposed to detect vehicles in night scenes. Firstly, an improved Generative Adversarial Network (GAN), named Attentive GAN, is used to enhance the vehicle features of nighttime images. Then, with the purpose of achieving a higher detection accuracy, a multiple local regression is employed in the regression branch, which predicts multiple bounding box offsets. An improved Region of Interest (RoI) pooling method is used to get distinguishing features in a classification branch based on Faster Region-based Convolutional Neural Network (R-CNN). Cross entropy loss is introduced to improve the accuracy of classification branch. The proposed method is examined with the proposed dataset, which is composed of the selected nighttime images from BDD-100k dataset (Berkeley Diverse Driving Database, including 100,000 images). Compared with a series of state-of-the-art detectors, the experiments demonstrate that the proposed algorithm can effectively contribute to vehicle detection accuracy in nighttime.

show abstract

“…Dark channel prior (DCP) techniques using physical phenomena, in which hues are degraded differently depending on the wavelength of the light, have been applied alongside general image processing algorithms in several stages, and mathematical model-based algorithms for underwater optical imaging processes have been proposed (He et al, 2010; al., 2012;Ancuti et al, 2017). Furthermore, deep learning techniques that are being successfully applied in the field of computer vision are being utilized to improve the quality of underwater images (Fabbri et al, 2018;Han et al, 2018;Uplavikar et al, 2019;Chen et al, 2019;Islam et al, 2020;Li and Cavallaro, 2020;Wang et al, 2020;Zhang et al, 2021). As the performance of deep learning techniques is critically affected by how the learning data is configured, underwater images and clean image data pairs are required to improve the quality of underwater images.…”

Section: Introductionmentioning

confidence: 99%

Comparison of GAN Deep Learning Methods for Underwater Optical Image Enhancement

Kim

Seo

Kim

2022

J. Ocean Eng. Technol.

View full text Add to dashboard Cite

Underwater optical images face various limitations that degrade the image quality compared with optical images taken in our atmosphere. Attenuation according to the wavelength of light and reflection by very small floating objects cause low contrast, blurry clarity, and color degradation in underwater images. We constructed an image data of the Korean sea and enhanced it by learning the characteristics of underwater images using the deep learning techniques of CycleGAN (cycle-consistent adversarial network), UGAN (underwater GAN), FUnIE-GAN (fast underwater image enhancement GAN). In addition, the underwater optical image was enhanced using the image processing technique of Image Fusion. For a quantitative performance comparison, UIQM (underwater image quality measure), which evaluates the performance of the enhancement in terms of colorfulness, sharpness, and contrast, and UCIQE (underwater color image quality evaluation), which evaluates the performance in terms of chroma, luminance, and saturation were calculated. For 100 underwater images taken in

show abstract

Underwater image enhancement using improved generative adversarial network

Cited by 10 publications

References 25 publications

Enhancement algorithm for high visibility of underwater images

Enhancement algorithm for high visibility of underwater images

A Nighttime Vehicle Detection Method with Attentive GAN for Accurate Classification and Regression

Comparison of GAN Deep Learning Methods for Underwater Optical Image Enhancement

Contact Info

Product

Resources

About