Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning

Yi Hung Sung; Li Wei Kang; Chia Hung Yeh

doi:10.1109/ICCE-Taiwan58799.2023.10226658

Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning

Yi Hung Sung^*
, Li Wei Kang^*
, Chia Hung Yeh

^*Corresponding author for this work

Department of Electrical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

2 Citations (Scopus)

Abstract

Underwater image restoration has gained more and more attention recently due to its several applications in marine environmental surveillance-related tasks. In this paper, a novel unsupervised GAN (generative adversarial network)-based deep learning framework for single underwater image restoration is proposed. Without needing paired training images, we introduce contrastive learning with feature and style reconstruction loss functions in our unsupervised GAN-based structure to learn an image generator for translating underwater images to the corresponding in-air images. Extensive experiments have shown that the proposed method outperforms (or is comparable with) the state-of-the-art deep learning-based methods relying on paired/unpaired training data quantitatively and qualitatively.

Original language	English
Title of host publication	2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	853-854
Number of pages	2
ISBN (Electronic)	9798350324174
DOIs	https://doi.org/10.1109/ICCE-Taiwan58799.2023.10226658
Publication status	Published - 2023
Event	2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Pingtung, Taiwan Duration: 2023 Jul 17 → 2023 Jul 19

Publication series

Name	2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings

Conference

Conference	2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023
Country/Territory	Taiwan
City	Pingtung
Period	2023/07/17 → 2023/07/19

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 14 Life Below Water

Keywords

contrastive learning
deep learning
generative adversarial networks
single underwater image restoration
unsupervised learning

ASJC Scopus subject areas

Artificial Intelligence
Human-Computer Interaction
Information Systems
Information Systems and Management
Electrical and Electronic Engineering
Media Technology
Instrumentation

Access to Document

10.1109/ICCE-Taiwan58799.2023.10226658

Cite this

Sung, Y. H., Kang, L. W., & Yeh, C. H. (2023). Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning. In 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings (pp. 853-854). (2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICCE-Taiwan58799.2023.10226658

Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning. / Sung, Yi Hung; Kang, Li Wei ; Yeh, Chia Hung.
2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2023. p. 853-854 (2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Sung, YH, Kang, LW & Yeh, CH 2023, Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning. in 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings. 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 853-854, 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023, Pingtung, Taiwan, 2023/07/17. https://doi.org/10.1109/ICCE-Taiwan58799.2023.10226658

Sung YH, Kang LW , Yeh CH. Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning. In 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2023. p. 853-854. (2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings). doi: 10.1109/ICCE-Taiwan58799.2023.10226658

Sung, Yi Hung ; Kang, Li Wei ; Yeh, Chia Hung. / Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning. 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2023. pp. 853-854 (2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings).

@inproceedings{fcb2449e31f24102ae127cb5968b6a5e,

title = "Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning",

abstract = "Underwater image restoration has gained more and more attention recently due to its several applications in marine environmental surveillance-related tasks. In this paper, a novel unsupervised GAN (generative adversarial network)-based deep learning framework for single underwater image restoration is proposed. Without needing paired training images, we introduce contrastive learning with feature and style reconstruction loss functions in our unsupervised GAN-based structure to learn an image generator for translating underwater images to the corresponding in-air images. Extensive experiments have shown that the proposed method outperforms (or is comparable with) the state-of-the-art deep learning-based methods relying on paired/unpaired training data quantitatively and qualitatively.",

keywords = "contrastive learning, deep learning, generative adversarial networks, single underwater image restoration, unsupervised learning",

author = "Sung, \{Yi Hung\} and Kang, \{Li Wei\} and Yeh, \{Chia Hung\}",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 ; Conference date: 17-07-2023 Through 19-07-2023",

year = "2023",

doi = "10.1109/ICCE-Taiwan58799.2023.10226658",

language = "English",

series = "2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "853--854",

booktitle = "2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings",

}

TY - GEN

T1 - Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning

AU - Sung, Yi Hung

AU - Kang, Li Wei

AU - Yeh, Chia Hung

PY - 2023

Y1 - 2023

N2 - Underwater image restoration has gained more and more attention recently due to its several applications in marine environmental surveillance-related tasks. In this paper, a novel unsupervised GAN (generative adversarial network)-based deep learning framework for single underwater image restoration is proposed. Without needing paired training images, we introduce contrastive learning with feature and style reconstruction loss functions in our unsupervised GAN-based structure to learn an image generator for translating underwater images to the corresponding in-air images. Extensive experiments have shown that the proposed method outperforms (or is comparable with) the state-of-the-art deep learning-based methods relying on paired/unpaired training data quantitatively and qualitatively.

AB - Underwater image restoration has gained more and more attention recently due to its several applications in marine environmental surveillance-related tasks. In this paper, a novel unsupervised GAN (generative adversarial network)-based deep learning framework for single underwater image restoration is proposed. Without needing paired training images, we introduce contrastive learning with feature and style reconstruction loss functions in our unsupervised GAN-based structure to learn an image generator for translating underwater images to the corresponding in-air images. Extensive experiments have shown that the proposed method outperforms (or is comparable with) the state-of-the-art deep learning-based methods relying on paired/unpaired training data quantitatively and qualitatively.

KW - contrastive learning

KW - deep learning

KW - generative adversarial networks

KW - single underwater image restoration

KW - unsupervised learning

UR - https://www.scopus.com/pages/publications/85174975994

UR - https://www.scopus.com/pages/publications/85174975994#tab=citedBy

U2 - 10.1109/ICCE-Taiwan58799.2023.10226658

DO - 10.1109/ICCE-Taiwan58799.2023.10226658

M3 - Conference contribution

AN - SCOPUS:85174975994

T3 - 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings

SP - 853

EP - 854

BT - 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2023 International Conference on Consumer Electronics - Taiwan, ICCE-Taiwan 2023

Y2 - 17 July 2023 through 19 July 2023

ER -

Single Underwater Image Restoration via Unsupervised Generative Adversarial Network and Contrastive Learning

Abstract

Publication series

Conference

UN SDGs

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this