基于密度-距离图的交互式体数据分类方法

发布时间：2017-12-31 01:16

本文关键词：基于密度-距离图的交互式体数据分类方法　出处：《软件学报》2016年05期 　论文类型：期刊论文

【摘要】：体数据分类是体绘制中传递函数设计的核心问题.标量值-梯度模直方图作为表征体数据的一种经典二维特征空间,已被广泛应用于分类体数据.然而,大部分已有方法存在过于依赖分类算法的参数设置、运算效率低、交互复杂度高等问题.以标量值-梯度模直方图的密度分布为基础,并依据物质中心密度大且物质中心间距离远这一特性,首先快速计算每个数据点的密度及每个数据点到比其密度大的点的最小距离;然后,将所有数据点投影到密度-距离图,并以密度-距离图作为人机接口,使用户能够交互地选择多个密度中心来分类体数据并设置传递函数.通过多组实验验证,所提出的方法无需预设物质类别的数量,分割标量值-梯度模直方图的准确度较高且速度较快,所设计的密度-距离图是一个有效的人机交互接口,可以有效地引导用户完成由粗糙到精细的递进式体数据分类和可视化过程.
[Abstract]:Volume data classification is the core problem of transfer function design in volume rendering. As a classical two-dimensional feature space to represent volume data scalar Value-gradient Modular histogram has been widely used in classifying volume data. Most of the existing methods are too dependent on the parameter setting of the classification algorithm, the operation efficiency is low, the interaction complexity is high, and so on, which is based on the density distribution of scalar value-gradient mode histogram. The density of each data point and the minimum distance from each data point to the point larger than the density of each data point are calculated quickly according to the characteristics of the density of the material center and the distance between the material center and the material center. Then, all the data points are projected to the density-distance graph, and the density-distance graph is used as the man-machine interface. It enables users to interactively select multiple density centers to classify volume data and set transfer functions. Through multiple experiments, the proposed method does not need to preset the number of material categories. The segmentation scalar value-gradient mode histogram has high accuracy and high speed. The density-distance map is an effective man-machine interface. It can effectively guide users to complete the process of classification and visualization of progressive volume data from rough to fine.
【作者单位】：中南大学信息科学与工程学院;中南大学软件学院;
【基金】：国家自然科学基金(61402540,61103108) 湖南省教育厅科学研究项目(13C095) 湖南省科技支撑计划(2014GK3049)~~
【分类号】：TP391.41
【正文快照】： Interactive Volume Data Classification Based on Density-Distance GraphZHOU Fang-Fang1,GAO Fei2,LIU Yong-Gang2,LIANG Xing1,ZHAO Ying11(School of Information Science and Engineering,Central South University,Changsha 410083,China)2(School of Software,Centra

【参考文献】