Surgical Scene Understanding Through a Geometric Lens
MCML Authors
Lennart Bastian
Dr.
Abstract
Lennart Bastian
Dr.
Abstract
This dissertation proposes advances in surgical scene understanding through geometric methods. Three core themes are addressed: (i) 3D scene understanding via various data modalities, (ii) geometric personnel modeling enabling staff-centric workflows, (iii) motion analysis for deformable and rigid objects. These techniques can potentially overcome 2D limitations, enabling robust semantic understanding, privacy-preserving tracking, and precise motion modeling for AI-driven surgical automation.
BibTeXKey: Bas25