Poster: A comparison of local shape descriptors for biological applications

3D imaging of biological structures is now ubiquitous. From these images it is possible to extract surface models of developing structures, gather examples of both healthy and diseased or injured populations, and track changes in structure due to disease over time. To make full use of these data sets we need tools to establish correspondences between surfaces with similar structures and quantitative measures of changes to local and global shape. A key requirement for both of these tasks is the ability to locally compare shape. More formally, the goal is to create a local shape descriptor which maps local shape to an n dimensional number. This descriptor can then be used for many downstream applications, such as comparing or quantifying shape change or automatically determining feature correspondences for registration.