基于掩模和非局部注意力的双阶段去雨网络

侯玉珍,沈晓红,李莉,杨明源,张彩明

Dual-stage deraining network based on mask and non-local attention

Yuzhen HOU,Xiaohong SHEN,Li LI,Mingyuan YANG,Caiming ZHANG

表 1 不同方法的客观评价指标对比

Tab.1 Comparison of objective evaluations of different methods

方法 Rain200L Rain200H DID-Data DDN-Data 平均指标 PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM PSNR/dB SSIM GMM (2015) 28.66 0.8652 14.50 0.4164 25.81 0.8344 27.55 0.8479 24.13 0.7410 DSC (2016) 27.16 0.8663 14.73 0.3815 24.24 0.8279 27.31 0.8373 23.36 0.7333 DDN (2017) 34.68 0.9671 26.05 0.8056 30.97 0.9116 30.00 0.9041 30.43 0.8971 PreNet (2019) 37.80 0.9814 29.04 0.8991 33.17 0.9481 32.60 0.9459 33.15 0.9436 RCDNet (2020) 39.17 0.9885 30.24 0.9048 34.08 0.9532 33.04 0.9472 34.13 0.9484 DualGCN (2021) 40.73 0.9886 31.15 0.9125 34.37 0.9620 33.01 0.9489 34.81 0.9530 SPDNet (2021) 40.50 0.9875 31.28 0.9207 34.57 0.9560 33.15 0.9457 34.88 0.9525 Restormer (2022) 40.99 0.9890 32.00 0.9344 35.29 0.9641 34.20 0.9571 35.62 0.9612 DRSgormer (2023) 41.23 0.9894 32.17 0.9326 35.35 0.9646 34.35 0.9588 35.78 0.9614 Regformer (2024) 41.51 0.9900 32.46 0.9353 35.43 0.9651 34.38 0.9591 35.95 0.9624 NeRD-Rain(2024) 41.71 0.9903 32.40 0.9373 35.53 0.9659 34.45 0.9596 36.02 0.9630 MMamba (2025) 41.49 0.9895 32.43 0.9345 35.41 0.9655 34.46 0.9593 35.95 0.9622 本研究算法 41.71 0.9907 33.23 0.9341 35.59 0.9680 34.54 0.9618 36.26 0.9636