search

menyifang / DCT-Net

Official implementation of "DCT-Net: Domain-Calibrated Translation for Portrait Stylization", SIGGRAPH 2022 (TOG); Multi-style cartoonization
Apache License 2.0
752 stars 75 forks source link

About the style representation in 3.4.1 #4

Open Leocien opened 2 years ago

Leocien commented 2 years ago

Hi, What is the difference between your style representation loss and the surface representation loss of white-box cartoonization?

menyifang commented 2 years ago

We followed white-box for style representation loss and no significant difference.

Leocien commented 2 years ago

We followed white-box for style representation loss and no significant difference.

As I know, Withe-box Method decomposes images into the surface representation, the structure representation, and the texture representation, which one is the style representation in your paper, or all of them?

menyifang commented 2 years ago

all of them are considered.