As part of the Visual Computing research colloquium we invite you to a new talk!

  • Date: Friday, 26.06.2026, 13:00
  • Room: G29-335
  • Speaker: Stefan Gumhold, TU Dresden

Efficient Visualization of Multivariate Line Data

Abstract:

Line data originates from a variety of scientific application domains and can represent characteristic lines in flow fields, fiber tracks reconstructed from diffusion tensor imaging, or trajectories of moving entities like humans, animals, vehicals, and cyber-physical systems. Line data is often 3D and decorated with multivariate data streams that can be sampled at different rates. This talk introduces efficient rendering techniques for 3D line data including support for transparent rendering as well as direct rendering of Hermite spline tube and ribbon segments. Furthermore, approaches to visualize multivariate data along 3D line data are presented. Finally, a user study is presented that compares tube vs ribbon based multivariate line data visualization. As a result advantages and disadvantages of both representation could be gathered for different visualization scenarios.

Biography:

Stefan Gumhold is a full professor at TU Dresden in the Faculty of Computer Science and heads the Chair of Computer Graphics and Visualization. His research interests span 3D geometry processing, scientific visualization, and scene understanding. Prof. Gumhold received his PhD in 2000 from the University of Tübingen and was postdoc in Craig Gotsman’s group at Technion in Israel as well as in the SCI-group at University of Utah. As a Heisenberg Scholar he visited the University of Granada in Spain and lead a research group on “3D Animation processing” at the Max-Planck Center for Visual Computing in Saarbrücken before he was appointed full professor at TU Dresden in 2005. At TU Dresden he served as dean of the Faculty of Computer Science and since 2018 chaired the environmental commission of the University, for which he received an honorary medal of the University in 2025. He has published more than 150 peer-reviewed publications.