基于改进CycleGAN的多失真类型水下图像增强

吕振鸣,董绍江,夏宗佑,牟小燕,王明权

Multi-distortion type underwater image enhancement based on improved CycleGAN

Zhenming LV,Shaojiang DONG,Zongyou XIA,Xiaoyan MOU,Mingquan WANG

表 1 几种算法在EUVP和UIEB中黄色失真图像上的定量评估结果

Tab.1 Quantitative evaluation results of several algorithms for yellow distorted images on EUVP and UIEB

算法 PSNR SSIM UIQM UCIQE CLAHE 19.6214 0.7557 2.8964 0.3396 UDCP 18.3616 0.7758 2.7784 0.3782 Retinex 12.6385 0.6869 2.4648 0.3629 IBLA 20.8086 0.8053 3.2923 0.4408 CycleGAN 17.9302 0.7431 3.4895 0.4025 FUnIE-GAN 19.4192 0.8163 3.2458 0.4496 LANet 18.6344 0.8255 3.2992 0.4507 SCEIR 22.9361 0.8482 3.4946 0.4673 MGCycleGAN 23.1957 0.8623 3.5961 0.4728