中文摘要：

目的：采用医学图像配准技术，对中国数字化可视人数据集进行准确、高效的配准。方法：①图像Lab色度空间变换和二值化；②图像分割和特征量提取；③构造坐标变换矩阵；④图像序列配准变换。结果：配准后的图像序列中，人体解剖结构在空间和结构上实现了精确的匹配对应。进行三维重建后，几何模型外表面具有较高的光滑度，说明方法有效、可靠。结论：采用特征提取和快速坐标变换的配准方法可较好地应用于数字化可视人数据集的图像配准。

英文摘要：

Objective：We apply medical image registration technology to the exact and efficient registration of the first dataset of Chinese digitized visible human. Methods：（1）The medical images were transformed in chrema space and 2-value. （2）We segmented the images and picked-up their eigenvectors. （3）The coordinate transformation matrix was built. （4）The image sequence was registered and transformed. Results： According to the registered image sequence, the exact anatomic matching in space and structure is realized. After 3D reconstruction, the surface of geometrical model is of higher glossy. It indicates that the method is effective and credible. Conclusion ：The registration method of picking-up eigenvector and fast coordinate transformation is more applicable to register the dataset of digitized visible human.