The paper mentions this portion here. I have 2 questions.
- Which layer are you extracting the full face from the eyes, and hair to get the face like the results above without eyes and hair.
- Which layer are you performing histogram matching from Y
I^ref
to I^src