SPADE论文笔记 Notes of SPADE thesis
Project Page：SPADE Project Page
Code: NVlabs/SPADE: Semantic Image Synthesis with SPADE
Paper：Semantic Image Synthesis with Spatially-Adaptive Normalization 基于空间自适应归一化的语义图像合成
NVIDIA的这篇文章将其算法称为SPADE，对应文章中所写到的Spatially-Adaptive Normalization*。但其本质其实是NVIDIA在2019年GTC中推出的GauGAN算法，也被人称为神笔马良，这一算法可以通过输入用户所画的简单轮廓图像，合成为真实图像，在合成的同时还可以选择合成图像的风格
This article by NVIDIA calls its algorithm SPADE, which corresponds to the *Spatially-Adaptive Normalization written in the article. But its essence is actually the GauGAN algorithm introduced by NVIDIA in GTC in 2019. It is also called the magic pen Ma Liang. This algorithm can synthesize simple contour image drawn by the user into a real image, and you can also select composite image style while synthesizing.