Abstract:  Over the last years, the fields of computer vision and computer graphics have increasingly converged. Using the exact same processes to model appearance during 3D reconstruction and rendering has shown tremendous benefits, especially when combined with machine learning techniques to model otherwise hard-to-capture or -simulate optical effects. In this talk, I will give an overview of some of my recent work in this area highlighting different representational choices and showing potential connections to other fields such as physical simulation. All of the presented methods aim at creating _interactive_ representations that can be rendered in real-time and interacted with using user input. While this is particularly challenging due to hard computational and quality constraints, I will show some interesting steps and insights working towards this goal.

Bio: Christoph Lassner is a Research Scientist at Epic Games where he is using techniques at the intersection of machine learning, computer vision and computer graphics to create new 3D reconstruction, rendering and animation techniques. He deeply cares about the computational challenges high quality reconstruction and rendering usually creates and has created the Pulsar differentiable renderer, now a rendering backend for PyTorch3D, and received an Honorable Mention at the ACM Multimedia Open-Source Software Competition for his contributions to open source. Before joining Epic Games, Christoph was leading a research team at Meta Reality Labs Research working on 3D reconstruction and neural rendering technology in the context of virtual humans. Before that, he worked at Bodylabs and Amazon on 3D and 2D human body pose and shape estimation technology. He received his PhD from the University of Tübingen, where he was a PhD student at the Bernstein Center for Computational Neuroscience and the Max-Planck Institute for Intelligent Systems.

Homepage:  https://christophlassner.de

Sponsored in part by:   Meta Reality Labs Pittsburgh