X-ray micro-tomography (a type of 3D microscopy) is a rapidly growing branch of computational imaging, where recent advances in imaging and computational hardware combine with improved algorithms to produce clear, high-resolution 3D models showing the interior of complex samples at the micron-scale. Tomography is, however, largely limited to the study of static samples that can remain entirely unchanging while views are collected from many hundreds of different angles. The ability to generate 3D models of evolving objects would reveal new information about many interesting systems in biology, the materials and earth sciences. 

Powerful ideas for the tomography of evolving objects have emerged in the past decade, with ANU playing a leading role, but these have remained largely theoretical.  This project will develop the most promising of these ideas from concept to reality, using ANU tomography instruments and advanced tomographic reconstruction toolkits. The project may form part of the ARC Training Center for Multiscale 3D Imaging, Modelling and Manufacturing, and will likely involve collaboration with the Australian Nuclear Science and Technology Organisation (ANSTO).

The project will focus on one of two categories of dynamic processes: fluid flow or mechanical deformation. Possible fluid-flow processes include drought-tolerant plants resisting dehydration under water stress, the flow of supercritical CO2 through water-filled porous rocks in geo-sequestration experiments. Possible mechanical deformation scenarios include: (i) load/compression testing of composite materials and 3D printed parts; and (ii) unavoidable deformation of non-rigid (e.g. biological) samples during scanning.

Both scenarios confound current tomographic imaging techniques since the static and dynamic aspects cannot be separated in individual projection views; any motion introduces motion blur and streaking in the reconstructed images. This project will develop a tomographic reconstruction method that incorporates a model for non-rigid motion, through: (1) a static reconstruction of the object at a specific instance in time, and (2) a 4D model of 3D deformation (motion map) over time.

Mathematics (Fourier transforms, Multivariable calculus, Linear algebra)
Physics (Optics)
Some experience in programming, preferably python.