BriefGPT.xyz
Apr, 2017
通过串联对抗网络实现轮廓上色
Outline Colorization through Tandem Adversarial Networks
HTML
PDF
Kevin Frans
TL;DR
本研究提出一种利用两个神经网络联合工作的方案，分别基于线条轮廓和彩色方案完成抠图和上色。通过限制传递的彩色方案信息进行泛化，最终实现了从零开始和混乱的彩色方案中进行抠图和上色的自然效果。
Abstract
When creating
digital art
,
coloring
and
shading
are often time consuming tasks that follow the same general patterns. A solution to automa
→