Quadric loss combined with Chamfer loss yields better sharp-feature reconstruction in 3D models than either loss alone or other point-surface alternatives.
Generating 3d faces using convolutional mesh autoencoders
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
-
Learning Embedding of 3D models with Quadric Loss
Quadric loss combined with Chamfer loss yields better sharp-feature reconstruction in 3D models than either loss alone or other point-surface alternatives.