Fragment-based Surface Rendering
Final Report Abstract
In retrospect, we were able to more or less exactly achieve what was outlined in the initial research proposal. We have shown that our research has significantly pushed forward the current state of the art in large model rendering, both in terms of memory requirements, quality and rendering speed. Unfortunately, we have to admit that the scientific output of this research project in terms of publications is rather low, even though we continually tried to publish our work. However, it was very shattering how aggressively reviewers in the field rejected this work from the top tier conferences. Even though we could clearly demonstrate significant improvements over existing work, the reviewers’ opinions were quite negative due to the engineeringlike aspects contained in our work. It was frustrating to see that today everyone seems to be talking about ”big data” and ”scalability”, but when researchers address these aspects from a more systems oriented point of view the acceptance goes down dramatically. In the future, we will in particular investigate whether the proposed data structure can also be used when local changes of the encoded model have to be updated, or additional models have to be merged with an already encoded model. Furthermore it will be investigated how to perform editing operations on a sample-based data structure, and how to extract high quality polygon representations from such a data structure efficiently. Building upon this research, a spin-off company was founded. It provides an interactive editing and rendering suite for 3D polygon data using voxel-based approaches.
Publications
- Hybrid Sample-based Surface Rendering, Workshop on Vision, Modeling and Visualization, 2012
Bürger, R. Westermann
- Visualization of Big SPH Simulations via Compressed Octree Grids, IEEE International Conference On BigData, 2013
F. Reichl, M.Treib, R. Westermann
