Multi-dimensional transfer function design based on flexible dimension projection embedded in parallel coordinates
Multi-dimensional transfer function design based on flexible dimension projection embedded in parallel coordinates
复制标题
DOI：
10.1109/pacificvis.2011.5742368
复制
发表时间：
2011-03
期刊：
影响因子：
--
通讯作者：
Hanqi Guo;He Xiao;Xiaoru Yuan
中科院分区：
文献类型：
--
作者：
Hanqi Guo;He Xiao;Xiaoru Yuan
在本文中，我们提出了一种有效的多元体积传递函数（TF）设计，提供了平行坐标图（PCP）、基于MDS的维度投影图和体积渲染图像空间的紧密耦合视图。在我们的设计中，显示每个变量维度的数据分布的PCP和显示降维特征的MDS无缝集成，为用户提供灵活的特征分类，而无需在不同数据呈现之间进行上下文切换。我们提出的界面使用户能够识别感兴趣的簇并使用套索、魔棒和其他工具分配光学属性。此外，还可以直接在体积渲染图像上绘制草图来探测和编辑特征。为了实现交互性，应用了高斯混合模型（GMM）的八叉树分区和其他数据缩减技术。我们的实验表明，所提出的方法对于多维 TF 设计和数据探索是有效的。
In this paper, we present an effective transfer function (TF) design for multivariate volume, providing tightly coupled views of parallel coordinates plot (PCP), MDS-based dimension projection plots, and volume rendered image space. In our design, the PCP showing the data distribution of each variate dimension and the MDS showing reduced dimensional features are integrated seamlessly to provide flexible feature classification for the user without context switching between different data presentations. Our proposed interface enables users to identify interested clusters and assign optical properties with lassos, magic wand and other tools. Furthermore, sketching directly on the volume rendered images has been implemented to probe and edit features. To achieve interactivity, octree partitioning with Gaussian Mixture Model (GMM), and other data reduction techniques are applied. Our experiments show that the proposed method is effective for multidimensional TF design and data exploration.