面向多天气退化图像恢复的自注意力扩散模型

Processing math: 100%

秦菁, 文渊博, 高涛, 刘瑶

A Transformer-Based Diffusion Model for All-in-One Weather-Degraded Image Restoration

QIN Jing, WEN Yuanbo, GAO Tao, LIU Yao

表2 不同算法在多天气退化图像恢复任务上的定量对比

Tab.2 Quantitative comparison of different methods in all-in-one weather-degraded image restoration tasks

方法	源	Snow100K-L PSNR/SSIM	Snow100K-M PSNR/SSIM	Snow100K-S PSNR/SSIM	Test1 PSNR/SSIM	Raindrop-A PSNR/SSIM
Uformer^[10]	CVPR 2022	26.24/0.8680	32.11/0.9316	34.00/0.9445	16.32/0.7565	30.33/0.9335
Restormer^[11]	CVPR 2022	$\underset{=}{29.57}$ / $\underset{=}{0.9106}$	$\underset{=}{33.71}$ / $\underset{=}{0.9490}$	$\underset{=}{35.43}$ / $\underset{=}{0.9583}$	$\underset{=}{29.81}$ / $\underset{=}{0.9208}$	$\underset{=}{31.10}$ / $\underset{=}{0.9337}$
All-in-One^[3]	CVPR 2020	28.14/0.8901	30.96/0.9290	32.63/0.9392	25.87/ $\underline{0.8996}$	$\underline{31.35}$ / $\underline{0.9299}$
TransWeather^[4]	CVPR 2022	$\underline{29.15}$ / $\underline{0.8930}$	$\underline{32.02}$ / $\underline{0.9343}$	$\underline{32.98}$ / $\underline{0.9447}$	$\underline{28.61}$ /0.8893	30.21/0.9179
AWIR-TDM		31.69/0.9240	35.47/0.9565	37.16/0.9642	31.68/0.9347	32.33/0.9429