The ANU CTLab houses several X-ray 3D microscopes, which are used to image samples for a wide variety of clients in industry and academia. These microscopes measure how the sample absorbs and refracts X-rays, and use this information to build detailed, geometrically-faithful 3D models of the sample's internal structure.

 

A small number of X-rays will "ricochet" or "scatter" through large angles as they interact with the sample. This is most significant when imaging very dense, or very large samples: e.g. metal aircraft parts, large 3D printed components, or samples imaged on the CTLab's new "whole core" scanner. If these scattered X-rays collide with the detector, their signal is largely hidden in the random measurement noise. However, because the scattered X-ray *do* possess an underlying structure, they lead to artefacts in the final 3D image of the sample.

 

The student will explore theoretical, numerical, and deep-learning based methods to model the underlying structure of the scatter, and develop methods to correct for its effects, both in-hardware (i.e. at the microscope) and in-software (i.e. during image analysis).

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