中文摘要：

为降低体绘制中传递函数参数选择的盲目性和设计的复杂性,提出一种基于拉普拉斯特征映射的传递函数设计方法。提取体数据中各种特征信息构建高维传递函数参数空间,通过拉普拉斯特征映射将其映射到保持了体数据局部流形结构和高维参数空间分类能力的二维参数空间,在此嵌入空间上设计一种基于k-means聚类的传递函数,得到了较好的体数据分类和绘制结果。通过在一组体数据上的实验验证了该方法的有效性。

英文摘要：

To reduce the blindness of parameter selection and the complexity of design when working on the transfer function during vo- lume rendering, a transfer function design method based on Laplacian eigenmap was proposed. First, various feature information was ex- tracted from volume data to form a high-dimensional parameter space. Then, this high dimensional parameter space was mapped to a two- dimensional space preserving the local manifold structure of volume data and the classification capabi-lity of the original space. Finally, a transfer function based on k-means clustering was designed in this embedded space and good volume classifying and rendering results were obtained. Several volume data sets were used in the experiment to verify the effectiveness of this method.