3D Image Segmentation and Shape Characterisation Driven by Topological Persistence

Tomographic imaging gives researchers the ability to view the interior of a tremendous variety of complex materials with striking clarity. Combining recent advances in the fields of image analysis and computational topology, we develop novel techniques for studying the structure contained within 3D tomographic images. These techniques enable the persistence of every individual topological feature to be scrutinised, leading to new classes of topology-driven image segmentation methods and structure characterisation tools. Together, these advances will result in significantly greater accuracy in the digital representation of physical objects, with important implications for the burgeoning fields of image-based simulation and prediction.