Ubiquitous volume rendering in the web platform

Résumé

The main thesis hypothesis is that ubiquitous volume rendering can be achieved using WebGL. The thesis enumerates the challenges that should be met to achieve that goal. The results allow web content developers the integration of interactive volume rendering within standard HTML5 web pages. Content developers only need to declare the X3D nodes that provide the rendering characteristics they desire. In contrast to the systems that provide specific GPU programs, the presented architecture creates automatically the GPU code required by the WebGL graphics pipeline. This code is generated directly from the X3D nodes declared in the virtual scene. Therefore, content developers do not need to know about the GPU.The thesis extends previous research on web compatible volume data structures for WebGL, ray-casting hybrid surface and volumetric rendering, progressive volume rendering and some specific problems related to the visualization of medical datasets. Finally, the thesis contributes to the X3D standard with some proposals to extend and improve the volume rendering component. The proposals are in an advance stage towards their acceptance by the Web3D Consortium.