At this year’s Central European Seminar on Computer Graphics (CESCG), Ole Siemers received the Best Paper Award for his work Image Based Level-of-Detail Optimization for Large-Scale 3D Reconstruction.
Ole’s work presents a method for improving the efficiency of large-scale 3D reconstruction based on Gaussian Splatting. Instead of optimizing a scene at a single resolution, the approach builds a hierarchy during training, allowing the reconstruction to adapt the amount of detail to the viewing conditions. This makes it possible to reduce unnecessary computation while preserving the visual quality needed for interactive exploration.
The paper is based on Ole’s Master’s thesis, supervised by Diana Marin and Peter Kán. It was developed in the context of the PostDisaster project, which investigates technologies for capturing, reconstructing, and exploring disaster-affected areas using drones, 3D reconstruction, and mixed reality. Within this setting, efficient scene representations are essential for turning large captured environments into usable visualizations for planning and decision-making.
CESCG provides a forum for students to present and discuss their research in computer graphics, visualization, and related fields. Receiving the Best Paper Award highlights both the quality of Ole’s work and its relevance for scalable 3D reconstruction in practical applications.
Congratulations to Ole on this excellent achievement!
Ole Siemers receiving the Best Paper Award at CESCG 2026.
Ole Siemers with Diana Marin, supervisor of the awarded paper.