Thomas Wimmer

This PhD project focuses on dynamic 3D scene understanding and generation, addressing challenges in real-time perception, interpretation, and synthesis of complex environments. The research explores novel approaches for efficiently capturing and modeling dynamic 3D scenes, integrating techniques from computer vision, machine learning, and geometry processing. One aim of the project can be learning of robust and powerful representations that can effectively capture the complexities and temporal coherence of (dynamic) 3D scenes, enabling multiple downstream usages in challenging environments. Possible goals of the project further include efficient ways of reconstructing dynamic 3D scenes with few or only single input cameras and in real-time.