Concepts of Hybrid Data Rendering

Torsten Gustafsson, Wito Engelke, Rickard Englund, Ingrid Hotz
Proceedings of SIGRAD 2017, Visual Computing, August 17-18, 2017, Norrköping, Sweden - 2017
Download the publication : 2017-Sigrad-HybridRendering.pdf [8.5Mo]  
We present a concept for interactive rendering of multiple data sets of varying type, including geometry and volumetric data, in one scene with correct transparency. Typical visualization applications involve multiple data fields from various sources. A thorough understanding of such data often requires combined rendering of theses fields. The choice of the visualization concepts, and thus the rendering techniques, depends on the context and type of the individual fields. Efficiently combining different techniques in one scene, however, is not always a straightforward task. We tackle this problem by using an A-buffer based approach to gather color and transparency information from different sources, combine them and generate the final output image. Thereby we put special emphasis on efficiency and low memory consumption to allow a smooth exploration of the data. Therefore, we compare different A-buffer implementations with respect to memory consumption and memory access pattern. Additionally we introduce an {early-fragment-discarding} heuristic using inter-frame information to speed up the rendering.

Images and movies


BibTex references

  author       = {Gustafsson, Torsten and Engelke, Wito and Englund, Rickard and Hotz, Ingrid},
  title        = {{Concepts of Hybrid Data Rendering}},
  booktitle    = {Proceedings of SIGRAD 2017, Visual Computing, August 17-18, 2017, Norrk{\"o}ping, Sweden},
  series       = {Link{\"o}ping Electronic Conference Proceedings},
  year         = {2017},
  publisher    = {Link{\"o}ping University Electronic Press, Link{\"o}pings universitet}

Author publication list