M.Sc. Domenico Riggio
Dr. Edoardo De Rose

Download als PDF

Motivation
In recent years, the need for computational tools to support planning, navigation, and surgical training in thoracic procedures has increased significantly. This project focuses on modelling lung collapse during surgical pneumothorax using collapsed CT data and finite element simulations to capture the underlying biomechanical behavior. The extracted simulation parameters are then applied to non-collapsed CT datasets to generate realistic synthetic collapsed anatomies. These synthetic data are used to train a deep learning model that can predict lung collapse directly from undeformed CT scans, reducing the need for time-consuming FEM simulations and supporting future AI-based tools for lung surgery.

Task
The project aims to explore GNN approaches to create reliable lung simulations that can enhance surgical planning and navigation.
1. Optimizing existing finite element collapse lung model
2. Data collection and dataset creation
3. Deep learning model development
4. Validation with real patient data

Requirements
• Python coding
• C++ coding
Good to have
• Mechanical modelling basics
• Deep learning fundamentals