About objective metrics: Is the SSIM can be predicted by PSNR ? (PSNR vs SSIM)

Question

I read some people saying that if you have a study that involve conclusion regarding PSNR you must repeat all again considering also SSIM.
Since this meaning double work, time (and why not money), doing all simulations again (that already got using PSNR) using the SSIM parameter now ?
Are SSIM and PSNR independent?

Answer 1

I have investigated this topic and found a good cited article:

icpr2010.org/pdfs/icpr2010_WeAT8.44.pdf‎

And the conclusion is: "As a final conclusion, it appears that the values of the PSNR can be predicted from the SSIM and vice-versa."

As well the article cite others article that saying PSNR and SSIM are not independent.