中文摘要：

传输函数设计是体数据可视化流程的重要环节,高效的传输函数设计方法是提升体数据可视化效率的关键.在传统传输函数设计的基础上,该文以提升体数据特征可视化及分析效率为目标,提出一种基于体数据空间相似性的传输函数优化设计方法.在特征空间分析过程中,结合体数据空间位置信息,定义特征空间相似性度量标准,对一维传输函数特征空间进行自适应划分;在光学参数映射过程中,定义能量方程描述感兴趣特征的可见性分布与目标可见性分布的差异,近似求解能量方程的梯度,加速光学参数向量的迭代优化,以高效地实现光学参数优化设计.相比于传统的传输函数设计方法,该方法有效耦合特征空间分析及光学参数映射过程,可以帮助用户快速地实现体数据中感兴趣特征的分析与可视化.大量的实验结果、效率对比及用户体验反馈信息进一步验证了该文算法的有效性与实用性.

英文摘要：

Transfer function is a key step in volume visualization, which also plays an important role in high-efficient volume exploration. Inspired by popular schemes, we propose a novel high-efficient transfer function, aiming at the effective combination of feature space analysis and optical parameter design. In the process of feature space analysis, a spatial similarity measurement is introduced to adaptively classify internal features in traditional 1D transfer function space, with the spatial information of internal voxels considered. Then, an energy equation is defined to depict the difference between current visibility distribution and the target visibility distribution for features of interest. In order to accelerate the process of optical parameter optimization, we approximate the gradient computation of the energy equation, which largely reduces the time consumption for iterative optimization of opacity vectors. Compared with traditional transfer functions, the proposed scheme highly integrates feature space analysis and optical parameter design, and can help users quickly explore features of interest. A large number of comparison results and a relative user study further demonstrate the effectiveness and application value of our high-efficient transfer function.