Presentation Title: Multiple Dataset Visualization (MDV) Framework For Scalar Volume Data

Committee:

• Dr. Bijaya B. Karki(Chair Professor)
• Dr. S. Sitharama Iyengar
• Dr. Jianhua Chen
• Dr. Brygg Ullmer
• Charles Delzell(Dean's Representative)

Date: March 30, 2009
Time: 10:00 AM
Location: Coates Hall,Room 256

Abstract:
Many applications require comparative analysis of multiple datasets representing different samples, different conditions, different time instants, or different views in order to develop a better understanding of the scientific problem/system under consideration. One effective approach for such analysis is visualization of the data. To give a different view on visualization, which normally deals with the graphical rendering of a single dataset, we introduce the phrase multiple dataset visualization (MDV) by which we mean that two or more datasets of a given type are rendered concurrently in the same visualization. In this PhD thesis, we propose a simple, innovative MDV framework, which deals with some fundamental issues that arise when several datasets are visualized together. It is shown that MDV is an important concept for the cases where it is not possible to make an inference based on one dataset, and comparisons between many datasets are required to reveal cross-correlations among them. Our MDV framework follows a multithreaded architecture which consists of three core components, data preparation/loading, visualization and rendering, each of which runs on a separate thread. The visualization module - the major focus of this study, currently deals with isosurface extraction and texture-based rendering techniques. For isosurface extraction, our all-in-memory approach keeps multiple datasets under consideration and the corresponding geometric data in the main memory. Alternatively, the only-polygons- or points-in-memory approach reduces the memory usage by keeping the scalar data in the memory only until the geometric data are extracted. To better address the issues related to storage and computation, we develop adaptive data coherency and multiresolution schemes. The inter-dataset coherency scheme exploits the similarities among datasets to approximate the isosurfaces from similar regions of other datasets using the already generated polygon data for one or more reference datasets. Only those non-reference data blocks (i.e., octree nodes), which differ from the reference blocks, are directly processed. Our intra/inter-dataset multiresolution scheme processes the selected portions of each data volume or the selected entire data volumes at high (original) resolution and renders the rest data at varying levels of reduced resolution. The graphics hardware-accelerated approaches adopted for MDV include volume clipping, isosurface extraction and volume rendering, which use 3D textures and advanced per fragment operations. Performance measurements were carried out by considering up to 64 set of scalar volume data with size of 256 x 256 x 256 in normal desktop environments. With appropriate user-defined threshold criteria, our data coherency, multiresolution, texture-based MDV techniques maintain a linear time-N relationship, improve the geometry generation and rendering time, and increase the maximum N that can be handled (N: the number of datasets). Finally, we justify the effectiveness and usefulness of the proposed MDV by visualizing 3D scalar data from parallel quantum mechanical simulations of materials. The data represents the electronic structures (i.e., electron charge density distributions) of magnesium oxide and magnesium silicate under different pressure and temperature conditions.

All are invited.