Postgraduate and honours research projects - projects supported by the Cooperative Research Centre for Biomedical Imaging Development (CRC-BID)

Projects listed on this page are affiliated with the CRC-BID. Students with interests aligned with these projects may also be interested in Detector Development or projects affiliated with the Centre of Excellence in Coherent X-ray Science.

For more information on projects, please contact the listed supervisor.

PEDRO - Pixelated Emission Detector for RadioisotOpes

Dr Chris Hall, Dr John Gillam, Dr Matt Dimmock

The Pixelated Emission Detector for RadioisotOpes (PEDRO) is a state-of-the-art hybrid emission imaging system, optimised for small animal models. It utilizes a Compton-camera stack geometry behind a coded aperture mask to maximize detection efficiency whilst maintaining the enhanced resolution of pinhole imaging. Students working under this project will experience both Monte-Carlo simulations and a current developmental research system and be able to validate their simulations against experimental data.

1. Gamma-ray Tracking for PEDRO (Hons/PhD)

Gamma-ray tracking algorithms enable successful event ordering and improve the quality of the reconstructed image. Students will be required to write code to reconstruct the path of the gamma rays as they scatter through the detector volume. They will assess the implication of their algorithm on the final reconstructed image.

2. The Monte-Carlo Investigation of Collimator Effects on the Point Spread Function (PSF) for PEDRO (Hons/PhD)

The shape of the collimator aperture directly affects the magnification and quality of the reconstructed image. Students will be required to extend GEANT4 code (C++) to investigate the geometry of the pinhole and the variation in Point Spread Function.

3. Image Reconstruction for Single Photon Emission Imaging (Final year projects/Hons/PhD)

The type of image reconstruction, whether analytic or iterative, will affect the final image produced by a system. Students will be required to write (edit) an image reconstruction algorithm and quantify the resultant image quality.

4. An Investigation of Detector Efficiencies and Cross-talk Effects for PEDRO (Final year projects/Hons)

The use of segmented detectors leads to the requirement of add-back spectra to reconstruct the incident energy. This leads to cross-talk and noise effects that affect the detection efficiency of the system. The student will be required write an analysis code (in the language of their choice) and quantify these factors to validate simulations against experimental data. The extent of the investigation would depend on the level of the project.

Tissue specific projection x-ray imaging

Professor Rob Lewis (Monash Centre for Synchrotron Science)

Using polychromic x-ray sources and an energy resolving position sensitive photon counter, a study will be made of the potential for clinical radiography using the energy of the transmitted x-ray beam to distinguish tissue type. The work will involve modelling the forward transforms of the x-ray field from the source, on paths through the specimen to the detector. The model will include considerations of both elastic and inelastic scattering in the object, intervening air, and the detector.

By using probability theory and iterative algorithms a most likely solution to the object tissue type and morphology will be generated by comparing the observed data to that from the model. The errors in the solution will be examined and explained with a view to minimisation to a level of clinical usefulness.

The detector to be used in this work is a state of the art hybrid pixel detector which has been recently purchased by the MCSS.

Tissue specific computed tomography

Professor Rob Lewis (Monash Centre for Synchrotron Science)

Using an energy resolving detector array it should be possible to obtain important clinical information from a computed tomography scan which is not possible with energy insensitive devices. This study will model the raw data production from a CT system having photon counting and energy resolving detector elements. Ways of using this additional information to enhance the morphological definition of regions of interest in the sample will be researched. A computer model of the CT scanner will be developed. The model will be used to compare generated data with that obtained from a prototype CT scanner set up in the laboratory, using a recently purchased energy resolving hybrid pixel detector. Demonstration images will be created with realistic samples. An assessment of the usefulness of the images will be researched.

Reducing dose in radiography using an energy resolving position sensitive photon detector array

Professor Rob Lewis (Monash Centre for Synchrotron Science)

It has been hypothesised that by suitably weighting the signal that a photon provides to an image by its energy, lower doses for clinical radiographical examination can be obtained. This study will develop the theory behind this idea and test out this novel imaging process using laboratory prototype equipment. Computer models of energy resolving x-ray radiography systems will be developed using sophisticated Monte Carlo codes. Images of test objects will be generated and compared to real data obtained in the laboratory. The model will be used to generate data on phantoms which will be used to assess the quality of the images in a realistic medical application.

Room temperature semiconductors

Professor Rob Lewis (Monash Centre for Synchrotron Science)

Novel semiconductors for use in radiographic x-ray detectors are being developed by various commercial companies around the world. These materials currently have a limit on the x-ray flux they can detect, before they lose linearity. This study will seek to understand the rate limitations of these materials, and provide potential ways of improving the limit so that they might be used in high rate clinical imaging systems.

Computer models of the charge transport in the material will be generated. These will be used to compare real data from devices fabricated and tested in the laboratory. Means of improving the rate performance will be tested by fabricating prototype detector elements.

Reconstruction of reduced-dose, variable-resolution tomographic images from asymmetric discrete projection data

Dr Imants Svalbe (School of Physics)

This project seeks to reconstruct CT images at medium resolution (sufficient to establish body landmarks and anatomic context) but that include a sub-region of high spatial resolution over a specified region of interest. This scheme aims to deliver the maximum amount of diagnostic information in a single image acquisition, whilst subjecting the patient to the lowest possible total radiation dose. Achieving this objective requires new reconstruction algorithms and enhanced CT data acquisition.

The traditional equal-angle stepped-projection scans, shown below in a), would be replaced by adaptive, asymmetric "mojette" projections, as depicted in b), to reconstruct a slice with locally-variable resolution, as depicted in c). The reconstruction technique exploits information interpolated from symmetric "ghost" projections that are induced in discrete Radon space by the asymmetrically acquired data.

a) Schematic diagram of a traditional, symmetric angle CT projection acquisition. b) Asymmetric projection acquisition. c) A targeted high resolution region of interest inside a medium resolution image. The thick lines represent sources of x-rays that form incident parallel beams to create the measured absorption profiles. In b), the density of projection sample points would vary with each projection angle.