Quantitative analysis of PSNR, SSIM and RMSE

RMSE  The smaller the better.

As a rule of thumb , so to speak RMSE Values in 0.2 ~ 0.5 Between , It shows that the model can predict the data more accurately . Besides , Corrected r Square is greater than 0.75 Is a good value to show the accuracy . In some cases , Adjusted r The square is 0.4 Or more is acceptable .

PSRN The bigger the better

PSNR higher than 40dB It shows that the image quality is excellent （ That is, very close to the original image ）,
stay 30—40dB It usually means that the image quality is good （ That is, distortion can be detected but acceptable ）,
stay 20—30dB Poor image quality ;
Last ,PSNR lower than 20dB The image is not acceptable

SSIM The bigger the better

SSIM Value range [0,1], The bigger the value is. , Indicates that the smaller the image distortion .