Output details

Building 3-D Statistical Shape Models by Direct Optimization

<26> Establishing a meaningful correspondence across a set of surfaces is essential when describing their variation. Automatic methods for this problem are significant, since objects of biomedical interest have real surfaces, and cannot be adequately described by a planar shape. Human annotation is often difficult or impossible due to lack of anatomical landmarks. This paper applies this automatic method to several real-world datasets, and shows quantitatively superior results to other approaches. In collaboration with spin-out Imorphics (http://www.imorphics.com/), this work led to methods that have been used by major pharmaceutical and orthopaedic companies across a range of disease/therapeutic areas.