中文摘要：

通过对医学体数据进行剖切，在方便观察其内部组织结构的同时，保留周围潜在的重要环境信息，可提供一个整体的情景图像。提出一种基于组织分割的体数据剖切方法，该方法关注于构成体数据的语义层，从医学诊断中的现实需求出发，以逻辑拆分来剖切体数据，以提供更加完整、直观的显示。将梯度矢量扩散法扩展到三维情形，同时提出三维梯度矢量扩散的体数据分割方法。在剖切方法中，首先运用三维梯度矢量扩散法来立体分割医学体数据中的组织，然后通过不同的空间传输函数来操作不同的组织，实现体数据的三维空间剖切体操作。实验结果表明，该方法能够有效去除遮挡数据以揭露内部信息，能保留相关环境信息来增加剖切的可理解性，可取得良好的虚拟剖切效果。

英文摘要：

Splitting medical volume data can provide the image with a whole scene by remaining potentially important surrounding contextual information while observing the internal structure of objects. A splitting method for volume data based on tissue segmentation was presented in this paper. The method paid special attention to the semantic layers composing a volume. Intact and intuitive visualization could be offered by logical splitting of volume data to meet the actual demands of medical diagnosis. In order to extent 2D gradient vector diffusion （ GVD ） method to 3 D case, a 3 D GVD scheme for volume data segmentation was proposed. 3 D GVD scheme was used to sterically segment tissues of medical volume data and then various spatial transfer functions were used to manipulate different tissues for splitting volume manipulation in 3 D space for volume data. Experiments showed that the occluding data was removed and information of internal side was effectively uncovered. Meanwhile,the surrounding contextual information was remained to improve the understandability of splitting.