AG3D: 从 2D 图像集合中学习生成 3D 头像（英文版）

英文标题：AG3D: Learning to Generate 3D Avatars from 2D Image Collections中文摘要：本文提出了一种从 2D 图像生成逼真的 3D 人物的新方法，并使用多个辨别器和 2D 法线图形式的几何线索来训练模型，取得了在几何和外观方面超过以前 3D 和关节感知方法的性能，通过系统的消融研究验证了模型的有效性和每个组件的重要性。英文摘要：While progress in 2D generative models of human appearance has been rapid,many applications require 3D avatars that can be animated and rendered.Unfortunately, most existing methods for learning generative models of 3Dhumans with diverse shape and appearance require 3D training data, which islimited and