双维度交叉融合驱动的图像超分辨率重建方法

贾晓芬,王子祥,赵佰亭,梁镇洹,胡锐

Image super-resolution reconstruction method driven by two-dimensional cross-fusion

Xiaofen JIA,Zixiang WANG,Baiting ZHAO,Zhenhuan LIANG,Rui HU

表 3 5种CGT退化网络的实验结果

Tab.3 Experimental results of five CGT degenerative networks

方法 基线模块 SIGA CCA CFB PM/106 FL/109 Set5 Set14 BSD100 Urban100 PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM CGT-S √ √ × × 11.21 52.82 32.61 0.9003 28.94 0.7899 27.79 0.7444 26.89 0.8105 CGT-SF √ √ × √ 11.21 53.44 32.62 0.9008 28.95 0.7902 27.80 0.7446 26.94 0.8114 CGT-C √ × √ × 11.20 47.61 32.52 0.8991 28.81 0.7865 27.71 0.7412 26.57 0.7996 CGT-CF √ × √ √ 11.20 48.23 32.53 0.8990 28.82 0.7871 27.72 0.7412 26.58 0.8006 CGT-SC √ √ √ × 11.21 50.22 32.63 0.9006 28.94 0.7897 27.80 0.7443 26.84 0.8081 CGT √ √ √ √ 11.21 50.83 32.81 0.9024 29.03 0.7921 27.85 0.7456 27.12 0.8155