基于峰值感知和多尺度约束的加权引导滤波器

张全,刘海忠

Weighted guided filter based on peak-aware and multi-scale constraints

Quan ZHANG,Haizhong LIU

表 3 不同滤波器模型的图像平滑结果对比

Tab.3 Comparison of image smoothing results for different filter models

$\lambda $ 模型 PSNR/dB SSIM $r = 2$ $r = 4$ $r = 8$ $r = 2$ $r = 4$ $r = 8$ 0.12 GIF 30.98 30.17 29.71 0.9041 0.8978 0.9076 WGIF 31.29 30.42 29.92 0.9063 0.8996 0.9092 GGIF 36.03 34.99 34.57 0.9495 0.9434 0.9485 SWGIF 37.32 34.47 32.43 0.9830 0.9641 0.9473 SKWGIF 32.24 31.22 30.79 0.9199 0.9047 0.9135 本研究 39.10 38.34 38.18 0.9763 0.9754 0.9627 0.22 GIF 26.85 25.26 23.90 0.8105 0.7649 0.7506 WGIF 27.26 25.59 24.14 0.8174 0.7721 0.7571 GGIF 34.14 32.05 30.47 0.9291 0.9071 0.8111 SWGIF 34.38 30.31 27.44 0.9731 0.9270 0.8758 SKWGIF 28.46 26.55 25.11 0.8518 0.7909 0.7718 本研究 36.23 34.64 33.68 0.9628 0.9511 0.9502 0.42 GIF 24.53 22.53 20.67 0.7368 0.6474 0.5860 WGIF 24.91 22.81 20.84 0.7451 0.6560 0.5934 GGIF 33.51 30.80 28.39 0.9208 0.8871 0.8663 SWGIF 32.96 28.02 24.58 0.9677 0.8961 0.8011 SKWGIF 26.28 23.83 21.78 0.7999 0.6921 0.6210 本研究 34.73 32.60 31.01 0.9533 0.9330 0.9239