将颅面模型数据分区问题转换为一种模式分类问题,给出了一种网格模型上多尺度的特征处理方法,提出了一种基于核方法的支持向量数据域描述(SVDD)数据分区方法。实验证明,该分区方法能快速、有效地对颅面模型的特征区域进行精确合理的分区,且能够适用于有复杂轮廓与形状的特征区域。
This paper refered to switch the craniofacial model data partition into a classified pattern,and proposed a data partitioning method based on kernel methods supporting by vector data description.Experiment shows that this method can be faster and more precise to get the partitioned characteristics,and be applied to identify the complex contours and shapes of feature regions.