PU-GAN is a GAN-based point cloud upsampling network using an up-down-up expansion unit, self-attention, and a compound adversarial-uniform-reconstruction loss to improve distribution uniformity and reconstruction quality.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
PU-GAN: a Point Cloud Upsampling Adversarial Network
PU-GAN is a GAN-based point cloud upsampling network using an up-down-up expansion unit, self-attention, and a compound adversarial-uniform-reconstruction loss to improve distribution uniformity and reconstruction quality.