Interactive Collision Detection for Molecular Visualization

Type: 
BA/PR
Persons: 
1

Description

Molecular scenes often consist of a large number of molecules, which move in what appears to be a random motion (Brownian motion). This introduces a high frequency of collision between individual molecules. For interactive applications, we need a high-performance implementation of the collision detection.
There are different data structures and techniques that can greatly improve the performance. For instance, one optimization technique is called “fast fixed-radius nearest neighbors” (http://on-demand.gputechconf.com/gtc/2014/presentations/S4117-fast-fixed-radius-nearest-neighbor-gpu.pdf). It greatly improves the computation time for neighbor search.

Tasks

The task of this work is to investigate, which collision detection technique is best suitable for molecular data and to implement it on the GPU. The final outcome of the project should showcase a dynamic molecular scene (see: https://www.youtube.com/watch?v=8IlxsEzTbxw).  

Requirements

  • Knowledge of English language (source code comments and final report should be in English)
  • Knowledge of GPU programming
  • Knowledge of C++ and Qt is advantageous

Environment

The project will be implemtented as an integration into an existing framwork. The framework is already capable of visualizing molecules.

Contact

For more information please contact Tobias Klein (tklein@cg.tuwien.ac.at).