基于均值聚类的体绘制中传递函数的自动产生

采用均值聚类的算法为三维数据场直接体绘制的传递函数中的各类分界自动产生关键点。该方法为传递函数关键点的初选提供了合理的依据和参考，将器官组织的灰度自动分为C类，传递函数的关键点取在边界点处，并使分界点处的透明度为较大值，这样使器官组织的分割自动达到较好状态。实验显示，通过自动产生的传递函数比手动选择的传递函数提供了表达更清晰组织的重构的显示效果。

Automatic Generating of Transfer Function Based on K-means Clustering in Direct Volume Rendering

The boundary key points of the transfer function curves are automatically generated by using k-means clustering algorithm for the three-dimensional data set in direct vohune rendering. It provides a reasonable basis and reference for the primary key points of the transfer function. In this method, the gray value of organs will be automatically divided into c classes by k-mean clustering, the key points of the transfer function will set in clustered border points, and the transparency of the cut-off points would be set up larger values. This will enablethe division of the organs to automatically achieve a better state. Experiments show that transfer function provides a more clear tissue reconstruction effect than the manual selection does.