gafniguy / 4D-Facial-Avatars

Dynamic Neural Radiance Fields for Monocular 4D Facial Avater Reconstruction

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Could not get the psnr result as paper mentioned

Rainiver opened this issue · comments

Hi, thanks for your great work ! I trained the model and could get visual good result.The frame results are attached below:
0185
I trained 40k iterations, however,the validation psnr is less than 20, far below of the result about psnr 26.85. I wonder if it is because of the different dataset. So could you please provide the dataset which can be tested psnr 26.85 in the paper ?

Hi,

The numbers are averaged over 3 test subjects. For this one, the PSNR over the test sequence is 21.8949. For the other two actors it's 30.3849 and 28.2824. I guess this one (or his shoulders) occupies more pixels in the image hence lower PSNR for this subject.

I've published recently the dataset for the other 2 subjects too.

Thanks for your reply. This really helps me!