Department of Computer Science, Michigan Technological University, Houghton, MI 49931, USA.
IEEE Trans Vis Comput Graph. 2013 Mar;19(3):393-406. doi: 10.1109/TVCG.2012.143.
We treat streamline selection and viewpoint selection as symmetric problems which are formulated into a unified information-theoretic framework. This is achieved by building two interrelated information channels between a pool of candidate streamlines and a set of sample viewpoints. We define the streamline information to select best streamlines and in a similar manner, define the viewpoint information to select best viewpoints. Furthermore, we propose solutions to streamline clustering and viewpoint partitioning based on the representativeness of streamlines and viewpoints, respectively. Finally, we define a camera path that passes through all selected viewpoints for automatic flow field exploration. We demonstrate the robustness of our approach by showing experimental results with different flow data sets, and conducting rigorous comparisons between our algorithm and other seed placement or streamline selection algorithms based on information theory.
我们将流线选择和视点选择视为对称问题，并将其形式化为统一的信息论框架。这是通过在候选流线的池和一组样本视点之间建立两个相关联的信息通道来实现的。我们定义了流线信息来选择最佳流线，并以类似的方式定义了视点信息来选择最佳视点。此外，我们分别基于流线和视点的代表性提出了流线聚类和视点分区的解决方案。最后，我们定义了一条穿过所有选定视点的相机路径，用于自动进行流场探索。我们通过展示不同流数据集的实验结果，并基于信息论对我们的算法与其他种子放置或流线选择算法进行严格比较，证明了我们方法的稳健性。
