NSF 0429753 Autostereoscopic Visualization and Geometric Computing for Biological Macromolecules
The planned research seeks to develop tools and techniques for visualization and geometric reasoning for large proteomics data. The proposed research has three main components -- implicit-function-based representations for proteins, faster computation of protein electrostatics, and real-time auto-stereoscopic visualization. The research will develop a unified framework of implicit functions that can compactly and efficiently represent non-bonded protein properties including protein electrostatics and accessibility. By tightly coupling these properties, implicit-function-based representations will likely result in synergistically favorable crossovers in the solutions to these problems. Further, providing adaptive, progressive, and hierarchical representations for implicit functions will allow them to be used in multiresolution schemes for storage, transmission, and computational steering. Protein electrostatic interactions are of central importance for many biological processes. Autostereoscopic visualization of proteins and their properties is important in understanding the 3D structure of proteins. The proposed research will develop new methods for rendering implicit surfaces and volumes using the recent advances in graphics hardware as well as adapt efficient methods for autostereoscopic displays with personalization.
