Smoothing twice about fastphotostyle HOT 3 CLOSED

Please check the experiment section for the following sentence

We also use the post-processing filtering step in Luan et al. [21] ...

The post-processing also smooths the image. That's why we call it smooth_filter.py in the code. This might be the source of confusion.

from fastphotostyle.

Check section 5 here in [21].

from fastphotostyle.

Thanks for your help, I had misunderstood. Mechrez et al. [22] experimented with training an image-to-image Conditional Generative Adversial Network to apply the same effect as their smoothing step.

I am interested to know if you tried anything similar? If not, do you have any views on that approach?

from fastphotostyle.