Table 1 The quantitative results of real-world denoising on AWLD dataset (PSNR/SSIM). Bolditalic number represents the best result, while bold number indicates the second-best result. F and P denote FLOPs and Parameters, respectively.
From: Wild horseshoe crab image denoising based on CNN-transformer architecture
Method | F(G) | P(M) | UI | UTI | TI | AVG |
|---|---|---|---|---|---|---|
AP-BSN20 | – | 3.7 | 35.37/0.9443 | 30.46/0.8026 | 29.05/0.8723 | 31.63/0.8731 |
DDT21 | 86 | 18.4 | 38.16/0.9606 | 32.54/0.9013 | 33.21/0.9143 | 34.64/0.9254 |
MIRNetv222 | 785 | 31.8 | 34.19/0.9361 | 31.91/0.8992 | 30.27/0.8836 | 32.12/0.9063 |
MM-BSN23 | – | 5.3 | 34.94/0.9344 | 31.23/0.8937 | 30.07/0.8821 | 32.08/0.9034 |
MPRNet24 | 573 | 15. 7 | 37.93/0.9528 | 31.94/0.9011 | 31.27/0.8914 | 33.71/0.9151 |
SwinIR25 | 18 | 5.17 | 37.64/0.9516 | 30.89/0.8875 | 33.04/0.9105 | 33.86/0.9165 |
DDPG26 | – | – | 30.13/0.8012 | 30.02/0.8801 | 26.31/0.8529 | 28.82/0.8447 |
Restormer27 | 155 | 26.1 | 37.59/0.9516 | 31.68/0.8915 | 32.15/0.9027 | 33.81/0.9153 |
ours | 196 | 28.3 | 38.12/0.9604 | 32.93/0.9026 | 33.59/0.9175 | 34.88/0.9273 |
