基于动态频域调制的交互式图像去雾网络

杨燕,宋鑫钰

Interactive image dehazing network based on dynamic frequency-domain modulation

Yan YANG,Xinyu SONG

表 1 不同去雾方法在各数据集上的PSNR和SSIM结果

Tab.1 PSNR and SSIM result of different dehazing algorithm on different dataset

方法 SOTS-outdoor NH-HAZE O-HAZE I-HAZE PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM GCA-Net[29] 31.98 0.9145 19.68 0.7017 21.63 0.7742 19.19 0.6529 Dehazeformer[14] 33.37 0.8928 22.07 0.8196 25.31 0.7837 20.67 0.6792 C2PNet[30] 32.27 0.8073 18.51 0.7416 20.45 0.7062 21.79 0.7285 DEA-Net[31] 33.53 0.9854 20.20 0.7615 22.73 0.7537 20.35 0.6944 DehazeXL[32] 28.56 0.8577 17.32 0.7060 24.95 0.8053 22.91 0.7630 本文方法 33.91 0.9631 23.71 0.8297 26.54 0.8211 24.93 0.8126