BriefGPT.xyz
Nov, 2020
将2D StyleGAN升级为3D感知人脸生成
Lifting 2D StyleGAN for 3D-Aware Face Generation
HTML
PDF
Yichun Shi, Divyansh Aggarwal, Anil K. Jain
TL;DR
提出LiftedGAN框架，无需手动标注或3DMM模型的训练，通过可微分的渲染器分解StyleGAN2的潜在空间，生成3D组件，并输出3D形状和纹理，实现内容可控性，在生成逼真高质量图像方面表现出超越现有方法的优点。
Abstract
We propose a framework, called
liftedgan
, that disentangles and lifts a pre-trained
stylegan2
for
3d-aware
face generation. Our model is "
→