Function description
Cancer is a leading cause of death worldwide. In many cancers, surgery, where the surgeon removes malignant tissue, plays a pivotal role. A significant development in surgical procedures is tracking surgical instruments in conjunction with preoperative MRI/CT imaging to guide the procedure, ensuring more accurate, safer, and less invasive procedures.
This project will develop novel deep learning-based image-guided surgery techniques for surgical procedures in the abdomen. In this region, interoperative motility, proximity to major vessels, and critical structures like nerves are important considerations for using image-guided techniques. Accurate localization of the major vessels, both preoperative and interoperative imaging, is vital to enable image-guided surgeries. Using our extensive dataset, you will develop models for accurate segmentation of the major vessels, ensuring precise geometrical results through the use and development of state-of-the-art geometric deep learning techniques. Since your models will be used directly during image-guided surgeries, (almost) real-time registration of preoperative imaging with three-dimensional interoperative imaging is vital. As such, accurate and fast mapping is of primary importance for these novel deep learning-based image-guided surgeries.