Biomedical Imaging

The aim of our research is to develop new methods and software for characterising diseases from medical images, and for the training and planning of computer assisted interventions. On-going projects cover a wide range of imaging modalities (Magnetic Resonance Imaging, Positron Emission Tomography, Computerised Tomography, ultrasound, optical imaging) and aim at fusing anatomical, functional and molecular information. Our approach is to embed into image processing algorithms expert knowledge from physiology, biology, pathology, and the physics of the imaging acquisition. We are developing novel shape, signal, and anatomical models at different levels: body, organ, and tissue. Main applications include surgical simulation, neuroimaging, radiotherapy planning and small animal imaging.

Publications for the biomedical imaging team.

	Knee Segmentation The goal of the Knee Segmentation project is the development of an automated 3D segmentation system for MR images of the knee. The primary aim of this system is to obtain accurate and robust segmentation results for the cartilages of the knee. This new computer assisted tool will help to diagnose and monitor individual patient, but also to provide quantitative information to assess efficacy of new therapies for on osteoarthritis. Learn More...
	Alzheimer's Disease The aim of the Alzheimer's disease project is to develop algorithms and software to process and analyse MRI and PET scans both qualitatively and quantitatively. It is part of a larger CSIRO and Federal Government project: the Australian Imaging Biomarkers and Lifestyle study. Structural information from MRI is combined with functional and molecular information from PET. Our goal is to provide a complete assessment of individual patients and a better characterisation of disease progression, benchmarked against healthy typical brains computed from large longitudinal study databases. Learn More...
	Colonoscopy Simulator The aim of this project is to build a surgical training system to train novice and to improve the skills of experienced gastroenterologists for performing colonoscopy. The colonoscopy simulator allows navigation through, and interaction with a 3D colon, generated from a large database of patient CT scans. It includes a haptic device for force feedback and adjustable realistic biomechanical models of the colon, whose complexity in geometry and pathology can be tailored to suit training needs. Learn More...
	Prostate Radiotherapy This project aims to develop methods to use high contrast MRI scans directly for prostate cancer radiotherapy treatment planning. A CT scan from the same patient will not be required, as the prostate will be automatically delineated and electron densities are assigned directly from MRI values. The main benefit will be to improve treatment outcomes by reducing radiation dosage to normal tissues, and increasing dosage to the prostate. Learn More...
	Characterization of new PET imaging agents Novel positron imaging contrast agents can be designed in large numbers to bind specific targets in vivo, and need to be screened and characterised. We are developing image processing methods to automatically segment organs from small rodents and estimate the main pharmaco-kinetics parameters of the new imaging compounds by using small animal SPECT, PET and CT systems.
	Cross Platform Medical Imaging Library Over the years we have implemented a large number of medical imaging algorithms under a common development platform: MILXTM. This allows us to develop new functions based on the existing code base in a time efficient manner, while enforcing a Software Engineering environment that ensures maintainable and tested code. The Cross Platform Medical Imaging Library (MILX™) expands on algorithms provided by the Insight Tool Kit (ITK), the Visualisation Tool Kit (VTK), and other open source libraries. Learn More...