Visualisation of multi-dimensional medical images with application to brain electrical impedance tomography

Medical imaging plays an important role in modem medicine. With the increasing complexity and information presented by medical images, visualisation is vital for medical research and clinical applications to interpret the information presented in these images. The aim of this research is to investigate improvements to medical image visualisation, particularly for multi-dimensional medical image datasets. A recently
developed medical imaging technique known as Electrical Impedance Tomography (EIT) is presented as a demonstration. To fulfil the aim, three main efforts are included in this work.
First, a novel scheme for the processmg of brain EIT data with SPM (Statistical Parametric Mapping) to detect ROI (Regions of Interest) in the data is proposed based on a theoretical analysis. To evaluate the feasibility of this scheme, two types of experiments are carried out: one is implemented with simulated EIT data, and the other is performed with human brain EIT data under visual stimulation. The experimental
results demonstrate that: SPM is able to localise the expected ROI in EIT data correctly; and it is reasonable to use the balloon hemodynamic change model to simulate the
impedance change during brain function activity.
Secondly, to deal with the absence of human morphology information in EIT visualisation, an innovative landmark-based registration scheme is developed to register brain EIT image with a standard anatomical brain atlas.
Finally, a new task typology model is derived for task exploration in medical image visualisation, and a task-based system development methodology is proposed for the visualisation of multi-dimensional medical images. As a case study, a prototype visualisation system, named EIT5DVis, has been developed, following this methodology. to visualise five-dimensional brain EIT data. The EIT5DVis system is able to accept visualisation tasks through a graphical user interface; apply appropriate methods to analyse tasks, which include the ROI detection approach and registration scheme mentioned in the preceding paragraphs; and produce various visualisations.

A thesis submitted to Middlesex University in partial fulfilment of the requirements for the degree of Doctor of Philosophy.