Abstract
One of the purposes of the emerging discipline, Computational Anatomy, is construction of Computational Anatomy Model, which is a set of computational and statistical expressions of anatomical knowledge including organ shapes and structures and plays a key role in medical image understanding. In addition, the model is finally mapped to new patient data for supporting diagnosis and therapy in clinical processes. That is, the mathematical problems on computational anatomy oriented to clinical applications lead to the subjects; how to express anatomical knowledge for medical image understanding and how to map them to new data. In this paper, we briefly review literature and discuss the subjects and also present our research results focusing on landmarks as a minimum unit for anatomical knowledge, model-image registration by using semi-landmarks and variational shape average of planar polygons for organ shapes.
