Hardware Accelerated Medical Data Visualisation on the Web

This paper proposes a way to develop high-quality real-time web-based medical 3D data visualisation. Isosurface extraction is used as an example to discuss how to use programmable Graphics Processing Units (GPUs) and shaders to improve rendering performance on the web. The method is designed to reduce data transmission. When data is ready, the performance penalty can be considered negligible. A method to estimate memory usage to balance client memory limitation and rendering quality is also described. A way of using the frame rate to measure performance on the web is suggested, which could be used in future web visualisation.