Open oranges01 opened 10 months ago
I had the same problem as you.I calculated the metric between the network output picture and the original picture after the mask, and the result was poor. I don't know if it was a mistake in calculating metric or something else. In what color space did the author calculate metric? RGB or something else?
Could you please provide a detailed procedure for quantitative evaluation? I have been trying to perform the evaluation in my own way, but the results I obtain are significantly different from those reported in the paper. It's quite clear that I must have been doing something wrong. Your guidance on the correct procedure would be greatly appreciated.