BriefGPT.xyz
Dec, 2022
3DHumanGAN: 面向照片般逼真的三维感知人体图像生成
3DHumanGAN: Towards Photo-Realistic 3D-Aware Human Image Generation
HTML
PDF
Zhuoqian Yang, Shikai Li, Wayne Wu, Bo Dai
TL;DR
本文提出了一种名为3DHumanGAN的生成对抗网络（GAN），用于合成全身人体的图像，具有在不同视角和姿势下一致的外观。该模型采用了2D卷积背骨由3D姿态映射网络调制的生成器架构，能够生成具有3D人体先验知识和一致性的逼真图像，并通过对抗学习从网络图片中学习而来。
Abstract
We present 3DHumanGAN, a
3d
-aware generative adversarial network (
gan
) that synthesizes images of full-body humans with consistent appearances under different view-angles and body-poses. To tackle the representat
→