Masterarbeit MSTR-2018-101

Zahn, Sebastian: Interactive ray tracing of solvent excluded surfaces.
Universität Stuttgart, Fakultät Informatik, Elektrotechnik und Informationstechnik, Masterarbeit Nr. 101 (2018).
109 Seiten, englisch.

Domain experts in fields concerned with the behavior of molecules, for example biochemists, employ simulations to study a molecule’s individual properties and mutual interactions with other molecules. To obtain an intuitive spatial understanding of the returned data of the simulations, various visualization techniques such as molecular surfaces can be applied on the data. The solvent excluded surface depicts the boundary between the molecule’s and a solvent’s occupied space and therefore the molecules accessibility for the solvent. Insight about a molecule’s potential for interaction such as reactions can be gained by studying the surface’s shape visually. Current implementations for the visualization of the surface usually utilize GPU ray casting to achieve the performance required to allow interactivity such as viewpoint changing. However, this makes implementation of physically motivated effects like ambient occlusion or global illumination difficult. If compute resources do not contain GPUs, which is often the case in compute clusters, expensive software rasterization has to be employed instead. As CPUs offer less parallelism compared to GPUs, overhead introduced by the overdraw of thousands of primitives should be avoided. To mitigate these issues, CPU visualization approaches resurfaced again in recent times. In this work, the solvent excluded surface is visualized interactively using the classic ray tracing approach within the OSPRay CPU ray tracing framework. The described implementation is able to compute and visualize the solvent excluded surface for datasets composed of millions of atoms. Additionally, the surface supports transparency rendering, which allows implementation of a cavity visualization method that uses ambient occlusion.

Volltext und
andere Links
Abteilung(en)Universität Stuttgart, Institut für Visualisierung und Interaktive Systeme, Visualisierung und Interaktive Systeme
BetreuerErtl, Prof. Thomas; Krone, Michael; Rau, Tobias
Eingabedatum19. Juni 2019
   Publ. Informatik