From 2D to 3D surgical instrument tracking: a technique based on intervals and geometric cues
摘要
Computer-assisted interventions for enhanced experience in minimally invasive surgery (MIS) require motion tracking of the surgical instruments. This work presents a computationally efficient and marker-free method for tracking and visualizing 3D movements of surgical instruments in minimally invasive surgeries (MIS). Unlike existing deep-learning or marker-based systems, the proposed approach derives 3D motion directly from 2D segmentation maps using geometric and kinematic relations. The framework requires no physical modification to the instruments and can serve as a plug-in for existing surgical video datasets. Simulation and experimental validation using a motion capture system demonstrate high accuracy (