BriefGPT.xyz
Nov, 2018
单视图三维重建的视角先验学习
Learning View Priors for Single-view 3D Reconstruction
HTML
PDF
Hiroharu Kato, Tatsuya Harada
TL;DR
该研究提出了一种训练鉴别器和重构器的方法，在重构器中使用先验知识来纠正未被观察到的视角，以便从任意视角重新构建形状。与目前最先进的方法相比，实验结果表明该方法在合成和自然图像数据集上表现更好。
Abstract
There is some ambiguity in the
3d shape
of an object when the number of observed views is small. Because of this ambiguity, although a 3D
object reconstructor
can be trained using a single view or a few views per
→