university projects

During my study of computer science, I was involved in many interesting smaller and larger projects. At all the time I tried to do as much computer graphics as possible, even though I enjoyed the other subjects too. I got the opportunity to learn from 3 different computer graphics professors and work together with some very talented people. The details gives you a list of the projects worth mentioning.


21. June 2001 This day I gave my presentation about texture mapping algorithms in scan line based rendering. My implementation showed affine, perspective correct uv-interpolation and planar mapping. All perspective correct approaches suffer from the division per pixel. I showed an optimization that uses the properties of the depth-iso-lines.
2. July 2001 The defense presentation of our practical team training for software design. The team of 5 combatants was ordered, to develop a management software for trade shows, using Java and a very uncomfortable framework. After several adoptions to make our ideas fit that framework, we developed a very good prototype of that software. Instead of showing the program live, I prerecorded short videos. Doing so saved me a lot of mistyping and enabled me to show the entire program in a very short timeframe.
11. Jan. 2002 Lightzone and myself held a presentation about physics in computer games. We gave an historical overview, what different kinds of physics were required in early games, compared to the current ones in 2002. The second part tried to figure out, what limits game physics and what probably will be done better in near future. The presentation also showed Havok and MathEngine physics middleware for games.
I did my one semester practical exercises in computer graphics. The exercise were split into some simple OpenGL demos and an animated 3ds max scene. I tried to build a virtual model of the Fröbel Arena and showed some LAN party action.
July - Nov.
I realized my preparation project for the final diploma thesis. Professor Deussen, the computer graphics professor, had the problem of managing a lot of plant models. All plant models were stored as txf-file, therefore I developed a TxfViewer program, that gives a nice and fast preview of them. My program featured shadow mapping and a per pixel Blinn lighting model, using C for Graphics and 2.0 fragment programs. At the lowest end I implemented an 7 pass OpenGL 1.1 algorithm that performs very nice shadow mapping.
Oct. 2005 -
30. May 2006
During that time I worked on my diploma thesis. My topic was to analyze different approaches to optimize space partitioning data structures for complex dynamic virtual scenes. To do that I was supposed to write a C++ framework and implement these approaches. The new ideas I had initially in mind, vaporized as soon as I tried to evaluate them against reality and existing approaches. Even the existing approaches were quite good in resisting to come together in one framework. The framework had to be capable of real-time rendering, collision detection and ray tracing in both 2D, 3D and probably in more dimensions. I could not really finish the entire project in time. The major part of my thesis therefore consists of evaluations of different non-commercial libraries that should have helped me, but for one or another reason did not.
If you are interested look here for more details.