Paper 13381-32
An inverse rendering framework for tomographic volumetric additive manufacturing
28 January 2025 • 2:20 PM - 2:40 PM PST | Moscone South, Room 155 (Upper Mezz)
Abstract
This presentation about Tomographic Volumetric Additive Manufacturing (TVAM) shows that advances in the computer graphics literature can be leveraged to build a more general optimization framework for TVAM pattern generation.
We show that obtaining suitable projection patterns can be formulated as an inverse light transport problem. We present Dr.TVAM, a new open-source framework that implements a physically-based differentiable renderer suited to TVAM, producing high-quality patterns. This approach can account for various printing process effects like refraction and scattering, and outperforms prior methods.
Presenter
Baptiste Nicolet
EPFL (Switzerland)
Baptiste Nicolet a fourth year computer science PhD student in the Realistic Graphics Lab (RGL) at EPFL, supervised by Wenzel Jakob. Prior to joining EPFL, he obtained a double master's degree from École polytechnique and Télécom Paris. He did several research internships at Inria and NVIDIA.
His research currently focuses on designing robust differentiable rendering algorithms to solve challenging inverse problems involving light transport.