BriefGPT.xyz
May, 2022
3DILG:用于三维生成建模的非规则潜在格
3DILG: Irregular Latent Grids for 3D Generative Modeling
HTML
PDF
Biao Zhang, Matthias Nießner, Peter Wonka
TL;DR
通过使用神经场,利用三维空间内不规则的栅格表征方法,提高了点云形状重构以及生成模型的精确性和质量, 在单张高分辨率图像、低分辨率图像和分类条件下的生成模型中,实现了有关3D形状建模的革新工作。
Abstract
We propose a new representation for encoding
3d shapes
as
neural fields
. The representation is designed to be compatible with the transformer architecture and to benefit both
→