Development of an Augmented Reality System for Arthoscopic Procedures
Arthroscopic surgery limits surgeons' spatial awareness when instrument tips pass outside the visible field. This project develops an Augmented Reality navigation system that tracks surgical instruments in real time, predicts occluded tip positions, and renders them as overlays on the live arthroscopic image. Key contributions include an improved calibration pipeline for oblique-viewing arthroscopes and a 6-DOF instrument pose estimation system, validated across two clinical scenarios in a pilot user study.
Autonomous endoscope holder for dual-access surgery
Endoscopic surgery reduces tissue damage, blood loss, and infection risk compared to open surgery, but confines surgeons to narrow working spaces with limited, tool-linked visibility, particularly in monoportal procedures. The emerging "uniportal+" proposes nesting a tool inside an endoscope in one portal, enabling bimanual manipulation with an expanded field of view. This concept remains unrealized in practice, however, since no surgeon can simultaneously control two tool-in-scope assemblies unaided. This project builds on a compliant, robot-mounted endoscope holder developed in a previous master's thesis, extending it into an autonomous system that controls the endoscope so the surgeon can focus on tool manipulation. The system will be validated using a phantom model in a surgeon user study assessing the feasibility of performing realistic surgical steps. This project will enable practical dual-access endoscopic manipulation in confined anatomical spaces.
Development of a Multi-Camera Perception Setup in Room-Scale IMMERSIVE 3D Environment
The IMMERSE theatre is a room-scale immersive environment with a high-resolution stereoscopic 3D projection surface for up to 40 participants. To enable natural interaction, the visual content should respond to the people in the room. This requires a reliable link between camera-based scene observation and real-time rendering. This semester, bachelor or master project focuses on implementing, calibrating, and systematically testing a multi-camera setup in the IMMERSE theatre. The setup will use suitable visible-light cameras and may be complemented by thermal cameras to improve observation of people under challenging illumination. A first representative dataset will be collected to document the operating conditions and support future development of tracking and perception methods.
Force Sensing for Endoscopic Tools in Orthopedic Surgery
Accurate regulation of tool–tissue interaction forces is essential in surgery to ensure a safe and efficient intervention. While force sensing has been widely studied in surgical applications, existing approaches often require tool modifications or are limited to low-force interactions with soft tissues, which are unsuitable for orthopedic applications. This project aims to estimate distal interaction forces in handheld orthopedic endoscopic tools without modifying the instrument or constraining the surgeon’s motion. A six-degree-of-freedom force sensor will be integrated at the tool's proximal end, and a calibration method will be developed to map proximal measurements to distal forces. The system will be designed, manufactured, and experimentally validated using a dedicated test bench. Finally, interaction forces will be recorded during ex vivo surgical procedures, enabling quantitative analysis of tool–tissue interactions.
Multiview 3D Reconstruction and 6DoF Pose Estimation of Articulated Surgical Tools for Open Surgery
This thesis aims to develop a multiview pipeline for 6DoF pose estimation of articulated surgical instruments using synchronized RGB video streams from Gopros and Fullframe mirrorless cameras. Current methods are predominantly monocular [1] and struggle with occlusions, articulation, and depth ambiguity. Our approach combines segmentation, reconstruction, and pose estimation into a unified system tailored for open surgery.

Powered by  SiROP - the academic career network