Neural Convolutional Surfaces
This work addresses shape representation for 3D modeling, offering improved compression and editing capabilities, but it appears incremental as it builds on existing neural methods.
The paper tackles the problem of disentangling fine local geometry from coarse global structures in 3D shape representation, achieving better neural shape compression than state-of-the-art methods and enabling manipulation and transfer of shape details.
This work is concerned with a representation of shapes that disentangles fine, local and possibly repeating geometry, from global, coarse structures. Achieving such disentanglement leads to two unrelated advantages: i) a significant compression in the number of parameters required to represent a given geometry; ii) the ability to manipulate either global geometry, or local details, without harming the other. At the core of our approach lies a novel pipeline and neural architecture, which are optimized to represent one specific atlas, representing one 3D surface. Our pipeline and architecture are designed so that disentanglement of global geometry from local details is accomplished through optimization, in a completely unsupervised manner. We show that this approach achieves better neural shape compression than the state of the art, as well as enabling manipulation and transfer of shape details. Project page at http://geometry.cs.ucl.ac.uk/projects/2022/cnnmaps/ .