Open smoreira00 opened 1 year ago
Thanks for your question. Actually, there was a similar discussion in a previous issue. I've included my relevant reply below.
In my implementation, I also evaluated those black pixels in order to keep a coherent manner to computing other metrics, like SSIM. Since SSIM is calculated on image windows, it is inconvenient to left out those masked areas. Thus, I simply take the masked images for evaluation in all metrics.
For more discussion, please refer to #11.
Thank you so much! Left another doubt in #11.
When measuring PSNR, SSIM, and LIPIPS, does it take into account the pixels we previously masked? Or does it discard those pixels? For example, in my images I have a black border that I'm masking, but I'm not sure if in the evaluation those pixels are considered or if it only considers the pixels that are not masked.