Rigid and Deformable Medical Image Registration

This project investigates rigid and deformable medical image registration methods for robotic ultrasound systems (RUSS). The goal is to align preoperative images, such as magnetic resonance imaging (MRI), with intraoperative ultrasound and to maintain accurate anatomical correspondence during robotic scanning and intervention. Rigid registration estimates global differences in position and orientation, supporting anatomical localization and initialization of the robotic system. This component has produced initial methods and experimental results. The ongoing deformable registration work focuses on compensating for local tissue deformation and motion caused by respiration, probe–tissue interaction, and changes between preoperative and intraoperative imaging. The developed methods will be evaluated using medical imaging data and integrated into the RUSS workflow to support ultrasound probe positioning, image-guided navigation, and robotic needle intervention. The overall aim is to improve the accuracy, robustness, and automation of robotic ultrasound procedures in realistic application scenarios.