HARDWARE-ACCELERATED DISTORTION FOR VOLUME VISUALIZATION IN MEDICINE
Marcelo Cohen, Ken Brodlie, Nick Phillips · 2005
Medical imaging applications frequently deal with large datasets, usually three dimensional data extracted from a CT or MRI scanner. Volume visualization techniques provide an effective way to visualize these data but sometimes there is too much to see at once on the display. Thus ideas from the Information Visualization field can be borrowed to help achieve a more effective understanding of that data, particularly the notion of focus and context based on spatial distortion. In order to obtain an acceptable performance, it is possible to use the graphics processor (GPU) to accelerate the distortion itself. I. INTRODUCTION Volume visualization is a well known technique for assisting the diagnosis of a number of diseases, by producing a 3D view of the results of medical scans. For example, in a neurological application one may want to visualize an aneurysm which is a condition that occurs when the tissue of a blood vessel is weak at a certain point and then suffers a dilation. This can be a very dangerous situation and even cause death if the aneurysm ruptures - hence the need to identify this problem as soon as possible. But an aneurysm can show up in a number of sizes, and sometimes can be hard to identify amongst the vessel network of the brain. It is quite important to be able to see this network in its entirety, but sometimes it is just impossible to visualize both the aneurysm and the vessels with enough detail. Thus this research has been motivated by the need to view brain aneurysms with good quality but at the same time still providing the user with a full view of the surrounding arteries. Ideally this should be done as quickly as possible - so another goal was to provide the surgeon with a tool capable of producing a 3D image nearly in real time. This paper presents the ideas which led to the implementation of such a prototype and also reports some results.