The Joint Medical Visualisation Project:

There is a common problem facing medical teaching - access to dissection is becoming more and more limited and there is currently a reducing number of anatomists, thus radiological data will, of necessity, become an important teaching tool for anatomy in the near future. The purpose of this project is to develop new tools, resources that make use of 3D stereoscopic visualisation and investigate the new methods of teaching and learning that these tools enable.

The proposed solution is a combination of elements: low cost stereoscopic visualisation systems, a powerful volume rendering tool (OsgVolume) and an easy to use authoring environment (Present3D) that enables course materials, lectures and individual learning programs to be created.

Rather than view MRI/CT scan data as a series of individual slices, a volumetric program processes those slices and combines to create a complete 3 dimensional volume which can then be manipulated interactively. The use of stereoscopic visualisation gives true depth perception essential for understanding complex 3d data and volumetric data benefits greatly from this.

To work with the volume data interactively and in realtime a range of tools are required, clipping planes, adjustment of transparency of particular colours or luminance values, selection tools and others, so the area of interest can then be extracted or viewed as a complete entity within the overall volume, dramatically improving understanding of the actual shape, form and interconnectivity of, for instance, a brain tumour. Other more specific tools will be required to meet demands of the wide range of practitioners who will benefit from this project, but the first aim is to concentrate on the requirements for medical teaching and research, with use for patient information systems probably following.

The visual quality of MRI/CT scans and Confocal Microscopy data may be variable and this can be dramatically improved by preprocessing techniques. For instance, the attached image (volume.jpg) has had normals calculated as part of the volume creation process, this produces apparent iso-surfaces and enables the light source to be moved interactively over the surface allowing fine detail to be revealed.

Current medical volume rendering programs are usually limited in the size of volume dataset and resolution of images that can be displayed. It is the aim of this project to enable much larger datasets and extremely high resolution images to be used and still be fully interactive on standard computer hardware. The ultimate goal is to be be able to handle the entire Visible Human 2 data where each slice can be 120,000 x 90,000 pixels. Where possible, this will be achieved by using already available and proven techniques, like the paging of high resolution imagery currently used in the OSG for terrain visualisation.

A proof of concept has been completed which enables example MRI or CT scan data to be processed to recreate a 3D volume data that can then be loaded in Present3D an easy to use, open source, realtime interactive stereoscopic visualisation authoring tool. The results of this have informed the technical requirements for the OsgVolume project. As the project has developed it has become clear that there are potentially three complementary joint projects:

2.1 OsgVolume:

The open source project to develop an advanced medical volume rendering core and specific medical tools that are built on it. By its nature this project will be incremental and modular.

2.1.1 Phase One :

An initial enabling project this will include the development of the OsgVolume core volumetric module, OSG loaders to directly handle the Dicom format from the various medical scanners and incorporation of simple (easy to use) volumetric tools. As a minimum this project would aim to provide the basic functionality required to enable OsgVolume to be used for teaching purposes. Implement the required application(s), developing incrementally to provide opportunities to test out work in progress and refine the requirements and deliverables of the application(s) within the time/financial constraints of the project.

2.1.2 Phase Two and beyond:

Ongoing development of OsgVolume to increase size of dataset and improve imaging techniques and tools available. It is already obvious that a number of medical research projects will benefit from the availability of OsgVolume and it is anticipated that many custom applications will be built on the foundations of the OsgVolume project. For instance, the Royal College of Surgeons has indicated that it considers a pre-operative planning tool to be highly desirable. At this point it is hoped that the project will start to take a life of its own with a growing community of developers contributing back to the core project to ensure its ongoing development as has happened with the OSG itself.

2.2 Medical Data Repository:

It is essential to have access to a shared resource of high quality anonomised volumetric, polygonal, point cloud, video and image data that can be used in the creation of teaching materials. Although a major feature of the overall project is the OsgVolume project and the collection of MRI, CT and Confocal Microscopy is key to the success of this, the huge benefits of stereo photography and stereo video should not be overlooked. Laser scans, 3D models, segmented from volume data, built by hand or exported from other medical software should also be included. It is critical that accurate and relevant metadata is associated with each dataset. This project could be undertaken by a single institution but would benefit enormously from becoming a national or international resource. It is anticipated that this would be an internet based project and would be jointly funded by the participants.

2.3 Courseware Materials:

The use of Present3D as an authoring environment does mean that lecturers can create their own course materials, however, it would make sense for the schools to collaborate on the development of core materials that set the standards and develop the new techniques that will ensure that maximum benefit is derived from this step change in medical education and that proper assessment and quality control procedures are put in place and feed back into the ongoing development of OsgVolume. A Centre of Excellence is an obvious goal for this project that then brands the courseware materials and distributes for worldwide use on a non-profit basis.