For blind guassion color image denoising on CBSD68, I wonder if the wrong picture was published. PSNR and SSIM calculated are:
(34.02061589141382, 0.9562660026199677)
(31.45404626915882, 0.9347315967521247)
(28.32975252659295, 0.8901727933655766).
In addition, my own test PSNR results on CBSD68 and Kodak are 0.01dB less than the published ones, which is understandable. But it's 0.03-0.05 less on McMaster.
For blind guassion color image denoising on CBSD68, I wonder if the wrong picture was published. PSNR and SSIM calculated are: (34.02061589141382, 0.9562660026199677) (31.45404626915882, 0.9347315967521247) (28.32975252659295, 0.8901727933655766).
In addition, my own test PSNR results on CBSD68 and Kodak are 0.01dB less than the published ones, which is understandable. But it's 0.03-0.05 less on McMaster.