On Expressive Shape Representations from Generative Models to 3D Deep Learning

Ilke Demir (GL)
DeepScale
Computer Science

This talk will discuss and present a collection of approaches to extract and utilize several shape representations. We will start with building hiearchical and parametrized shape representations (grammars) from several shape formats, then explore their use cases for generative models, and continue with how similar constructions can be exploited in extracting shape abstractions in the context of 3D deep learning.


Back to Geometry and Learning from Data in 3D and Beyond