Poor Man’s Rendering Of Segmented Data

In this paper we present a set of techniques for fast and efficient rendering of segmented data. Our approach utilizes the expected difference between two co-located texture lookups of a label volume, taken with different interpolation filters, as a feature boundary indicator. This allows us to achieve smooth class boundaries without needing to explicitly sample all eight neighbors in the label volume as is the case with previous methods. We also present a data encoding scheme that greatly simplifies transfer function construction.

@inproceedings{LB13,
  author       = {Lindholm, Stefan and Bock, Alexander},
  title        = {{Poor Man’s Rendering Of Segmented Data}},
  booktitle    = {Proceedings of SIGRAD 2013, Visual Computing, June 13-14, 2013, Norrk{\"o}ping, Sweden},
  volume       = {094},
  pages        = {49--54},
  year         = {2013},
  editor       = {Timo Ropinski and Jonas Unger},
  publisher    = {Link{\"o}ping University Electronic Press}
}