The ability to visualise 3D medical images is key

Document Sample
The ability to visualise 3D medical images is key Powered By Docstoc
                                                                            Collaborative Medical Image Visualisation

“The ability to visualise                     This demonstrator shows how the Grid can be exploited to construct
3D medical images is key                      collaborative virtual environments for visualisation of large volumes
to the diagnosis of                           of data. It uses data acquired from medical imaging scans, such as MR
                                              (magnetic resonance) and PET (positron emission tomography)
pathologies and pre-
                                              scans. Multiple participants inhabit a virtual space that allows
surgical planning”                            interactive examination and remote consultation over data sets
                                              acquired from the human brain.

                                              The ability to visualise 3D medical images is key to the diagnosis of
                                              pathologies and pre-surgical planning, as well as to enhance the basic
                                              understanding of anatomy. Data visualisation aims to use computer
                                              graphics tools in order to interpret masses of data. An effective
                                              visualisation system is one with the ability to highlight key features
                                                           within large data sets and to present these to users in clear and
                                                           meaningful ways. With human brain data for example,
                                                           users need to see major anatomical features of the brain for
                                                           bearings and may also need to be shown anomalies such
                                                           as tumours or other signs of disease. Crucially, however,
                                                           the users must have the capability to investigate the full
                                                           richness of the data within particular regions of interest
                                                           they select. Collaborative visualisation allows multiple
                                                           users to view and interact with the data simultaneously,
                                                           allowing experts to be consulted while visualising the data.
                                                           The true power of the system demonstrated lies in the fact
                                                           that interactive consultations can be made between
                                                           collaborators in remote physical locations.

                                                           This demonstrator shows how Grid technology can be
                                                           combined with a state-of-the-art graphical display, ReaCTor, an
                                                           Immersive Projection Technology (IPT) environment developed
                                                           at UCL, in order to visualise this kind of data effectively. The
The light-grey mass visible on the right of   ReaCTor is a four-sided device, with three walls and a floor. Data
this MR image indicates abnormal tissue
signals caused by a brain tumour. The         visualisation is presented in head-tracked stereo in real-time,
white lines highlight neural fibre pathways   producing the illusion of 3D objects appearing both within and beyond the
passing through the tumour. Surgeons
must avoid damaging these fibres if they
are to remove the tumour without causing
death or disability.
                                              The process of extracting meaningful features from large data sets
                                              often requires lengthy computations, which prevent real-time
                                              interaction with the data. However, these processes are often
                                              parallelisable and can be performed in real-time if sufficient
                                              resources are available. Access to the Grid means these real-time
                                              computations can be carried out without the need for large on-site
                                              computational resources. Use of the Grid in this way allows the
                                              computation to be truly user-steered towards particular data
                                              processing tasks in interesting or anomalous regions of the brain.
                                              The immersed user is provided with a variety of configuration and
                                              exploration tools and, utilising Grid technology, can collaborate
A user at the UCL ReaCTor configuring the     with colleagues in other facilities, using either other IPTs or desktop
visualisation system.                         systems.

                                              In particular, the project demonstrates the following e-science
                                              technologies: access to a computational grid service that is
                                              distributed and massively multi-processor; sharing of the
                                              computational results between multiple participants who are
                                              participating in a synchronous session; access to data visualisation
                                              and multimedia collaboration tools within a unified virtual

                                              The demonstrator has given rise to novel research in medical
                                              imaging and image processing and in the design of algorithms to
                                              extract meaningful features from the data. The design of future
                                              human computer interfaces for virtual environments is an active
                                              area of research and valuable information on the behaviour of
                                              collaborating users within the shared spaces has also been gleaned
                                              from work on this demonstrator. Very positive feedback has been
A desktop user interacts with a user in the   received from neurologists and neurosurgeons who see great
ReaCTor                                       potential, both in the use of the system as a tool for consultation
                                              and surgical planning and also as an anatomical teaching aid.

                                                                                              For further information contact:
                                                                                   Dr Anthony Steed at
                                                                           Or Dr Daniel Alexander at