Underwater Image Enhancement Algorithm Using Color Correction and Attention Mechanism

Ziqian Peng; Zhiyuan Cheng; Jiahao Hu

doi:10.3788/LOP242255

Laser & Optoelectronics Progress, Volume. 62, Issue 14, 1401003(2025)

Underwater Image Enhancement Algorithm Using Color Correction and Attention Mechanism

Ziqian Peng^1,2, Zhiyuan Cheng^1,3、*, and Jiahao Hu^1,2

¹Photoelectric Tracking and Measurement Technology Laboratory, Xi'an Institute of Optics and Precision Mechanics of Chinese Academy of Sciences, Xi'an 710119, Shaanxi , China

²University of Chinese Academy of Sciences, Beijing 100049, China

³Key Laboratory of Space Precision Measurement Technology, Chinese Academy of Sciences, Xi'an 710119, Shaanxi , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(10)

Fig. 1. Jaffe McGlamey's underwater imaging physical model

Download full size

Fig. 2. UCM-GAN structure. (a) Training module; (b) enhanced image generation module

Download full size

Fig. 3. Schematic diagram of convolutional attention module

Download full size

Fig. 4. Schematic diagram of the combination of channel attention module and spatial attention module

Download full size

Fig. 5. Comparison between ReLU function and SiLU function

Download full size

Fig. 6. Comparison between the original images and the enhanced images

Download full size

Table 1. Comparison of PSNR, SSIM, and UIQM results of different algorithms on UIEB dataset

View table

Table 1. Comparison of PSNR, SSIM, and UIQM results of different algorithms on UIEB dataset

Algorithm	SSIM	PSNR	UIQM
UDCP^［21］	0.4204	10.3974	1.5854
UCM^［22］	0.7675	17.3677	2.4873
WaterNet^［23］	0.7719	15.9199	2.7335
UWCNN^［24］	0.6822	12.2580	2.6430
MBANet^［25］	0.8243	18.9268	2.9862
TOPAL^［26］	0.8729	19.8501	2.7679
Fuine-GAN^［27］	0.7309	16.8918	2.9766
USUIR^［28］	0.8480	19.7589	0.9415
CycleGAN^［29］	0.6379	15.5718	2.7647
UCM-GAN	0.8818	28.5890	3.1088

Table 2. Enhancement effect comparison
View table
Table 2. Enhancement effect comparison
Algorithm SSIM PSNR
MMLE^［31］ 0.7252 17.70
HLRP^［32］ 0.1929 12.64
UGAN^［33］ 0.7545 22.79
UCM-GAN 0.8242 23.18

Table 3. Comparison of loss function ablation experiments

View table

Table 3. Comparison of loss function ablation experiments

Loss function	PSNR	SSIM	UIQM
MSE	22.226	0.816	3.0905
Cross entropy	22.074	0.838	2.783
Loss of reference ［34］	23.066	0.837	2.828
Charles loss	21.0503	0.815	2.845
Modified loss of reference ［34］	24.114	0.6771	2.7802
Proposed loss	28.5890	0.8818	3.1088

Table 4. Comparison of preprocessing module ablation experiments
View table
Table 4. Comparison of preprocessing module ablation experiments
Preprocessing module PSNR SSIM UIQM
NON-GAN 19.6382 0.6174 2.8062
UCM-GAN 28.5890 0.8818 3.1088

Tools

Get Citation

Copy Citation Text

Ziqian Peng, Zhiyuan Cheng, Jiahao Hu. Underwater Image Enhancement Algorithm Using Color Correction and Attention Mechanism[J]. Laser & Optoelectronics Progress, 2025, 62(14): 1401003

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites