Background

In surgical interventions, surgeons manipulate soft tissue to remove cancerous regions and reconnect healthy structures into a functional state. A key aspect of this process is an intuitive understanding of tissue tension distribution, which helps avoid tissue damage and rupture.

In robotic surgery, such tension information is equally critical but very challenging to obtain directly. Visualizing tissue deformation offers a promising way to infer tension indirectly. Recent advances in any-point tracking methods provide an opportunity to bridge the gap between simulation and real-world data, enabling new research directions in surgical robotics.

Task

The work will start with the development of a simulation-based dataset, followed by a comparison of simulators and an approximation of the simulator results using a Graph Neural Network (GNN), which introduces a crucial speed-up compared to an actual physical simulator.

Subtasks

Within the scope of the thesis, the following points should be addressed:

• Literature review on soft-tissue biomechanics, tension estimation, and visual tracking in surgical robotics
• Construction of a simulation-based dataset capturing tissue deformations under varying conditions
• Development of methods (e.g., Graph Neural Networks) to estimate or approximate tissue tension from deformation information
• Quantitative evaluation in simulation
• Discussion of limitations and potential extensions towards clinical applicability

Literature

• ICRA 2025 Workshop Spotlight (JIGGLE)
• arXiv:2409.14282
• arXiv:2309.11655
• arXiv:2309.11656

Your Profile

• Studies in a field relevant to the task description (e.g., computer science)
• Strong background in machine learning and computer vision
• Proficiency in Python and deep learning frameworks such as PyTorch
• Interest in medicine and understanding of surgical robotics

Application

Please send your application with the listed attachments to one of the following addresses: dennis.schneider@tum.de, karaoglu@imfusion.com

Attachments: CV, Transcript of Records, Brief statement of motivation (including relevant background)