Development of a semi-automated segmentation framework for thoracic-abdominal organs

Due to the increasing amount of data available from medical imaging procedures and also the increase in computing power, there has been a rise in the automation of the analysis of such data. A crucial step in the automation of such procedures is accurate segmentation of anatomy. Popular approaches include model based segmentation. However, these approaches require an atlas which may not be generic enough. This paper proposes a semi-automated data-driven segmentation framework of thoracic CT scans. The preliminary results of the framework is presented and discussed with proposals for future work.