Manipulating Geometry through Deep Learning

Abstract:

Manipulation and deformation of 3D shapes is one of the most basic modes of interaction and creation in the 3D domain, for instance - 3D design, modeling, animation, and registration. Integration with deep learning has the potential to revolutionize various applications that depend on such tasks, e.g., generative techniques for fully functional 3D objects. In this talk I will discuss my line of work which aims to provide deep learning with capabilities for shape manipulation, focusing on my most recent ones which tackle applying deep learning to injective deformations.

Short-bio:

Noam Aigerman is an Assistant Professor at the University of Montreal and an Associate Member of Mila. Before that, he was a Research Scientist at Adobe Research. His work lies at the intersection of 3D geometry processing and deep learning and focuses on devising fundamental geometric approaches to ML-related problems in 3D.