目前颅面重构技术往往利用目标颅骨所反映出的种族、性别、年龄、营养情况等信息来选择参考颅骨或者参考面部模板进行颅面重构,但这些信息并不包括颅骨自身的信息(如形状信息);在得到颅面重构的结果后,评估生成的人脸到底像不像其本人的真实人脸,数值化评估重构结果尚待完善。基于以上问题,提出了一种颅面相似度比较的方法,将颅面特征点分为形状特征点和一般特征点,针对不同特征点计算不同距离来衡量颅面相似度。实验结果表明,利用该方法选择出的参考颅骨或面部模板与志愿者的选择基本一致,可用于进行颅面重构后续计算并对重构结果进行评估。
Current methods of craniofacial reconstruction always choose the reference skull or the facial template based on the values of race,gender,age,nutrition,but these values do not include the properties of the skull (shape value e. g. ) . In addition,whether the reconstruction resembles the person of the object skull or not is still a difficult problem. Therefore,this paper presented a method of craniofacial similarity comparison,in which classified landmarks into two categories and computed different distances,so the reference skull or facial template could be chosen for further computation and to estimate the final result. Experimental results demonstrate the results computed are consistent with what the volunteers chose.