A Fully Automated Approach to Segmentation and Registration of 3D Medical Image Data for Pulmonary Diagnosis

Molecular imaging is an important tool that has found wide-spread use in the diagnosis and observation of various diseases and has more recently been used in areas such as drug development in order to facilitate the observation and analysis of newly developed drugs. The amounts of data in drug development experiments may be very large due to the involvement of both spatial and temporal information of med-ical images. Imaging techniques can facilitate the analysis of this data by automating information extraction and providing meaningful results. We propose a fully automated approach to pulmonary diagnosis in the context of drug development experiments using image segmentation and registration techniques. In particular, we propose a modification of the Chan-Vese approach for a stable segmentation of the lungs which then serves as a starting point in a spatial alignment process. Our results demonstrate the feasibility and potential of the proposed approach.