Dealing with Massive Volumetric Visualization: Progressive Algorithms and Data Structures

Massive data sets coming from complex simulations are “unwill-ing ” to fit completely inside computer’s internal memory. Con-sequently it is increasingly more difficult to visualize simulations results interactively. The use of smart External Memory (EM) al-gorithm becomes a must. To address these issues we propose an elegant and simple to implement framework for performing out-of-core visualization and view dependent refinement of large vol-ume datasets. We adopt a method for view-dependent refinement that relies on a well known strategy (longest edge-bisection) yet introducing a new method for extending the technique to the field of Volume Visualization. We perform a top-down traversal of the mesh hierarchy to produce a continuous surface guaranteeing con-nectivity of the resulting subdivision and achieving good memory coherency in data accesses. We show how this framework extends the framework for terrain visualization of Pascucci and Lindstrom [Pascucci and Lindstrom 2001], to Volume Visualization still main-taining simplicity and memory/computing efficiency of the original technique.