BriefGPT.xyz
Jun, 2021
具有隐式位移场的几何一致性神经形状表示
Geometry-Consistent Neural Shape Representation with Implicit Displacement Fields
HTML
PDF
Wang Yifan, Lukas Rahmann, Olga Sorkine-Hornung
TL;DR
本文介绍了一种新颖的隐式位移场在3D几何形状的表达及重构中的应用，它将一个复杂的表面表示为一个平滑的基础表面及一个沿着基础表面法线方向的位移，实现了高低频信号的分解，提高了表达能力、训练稳定性及泛化性。
Abstract
We present
implicit displacement fields
, a novel representation for detailed
3d geometry
. Inspired by a classic surface deformation technique,
di
→