基于分区的破损颅骨性别判别方法

颅骨性别鉴定是法医人类学在进行身份认证中首要而重要的一步,在实际应用中破损颅骨性别鉴定更具有研究价值。为鉴定破损颅骨性别,缩小刑事案件搜索范围,提出一种不完整颅骨性别判别模型。首先,将颅骨划分为七个分区,标记特征点,量化不可测量特征;然后,利用基于最大似然估计的前向逐步回归方法选择每个分区的最优特征子集,通过logistic回归建立七个分区的性别判别模型,并通过留一法进行验证;最后,实现了女性和男性不完整颅骨的最终性别判别。实验结果表明,各个分区都可以确定颅骨性别,分区数组合越多,判别准确率越高。

Sex determination of incomplete skull based on partition

Skull gender identification is the first and important step in forensic anthropology authentication. In practice, the incomplete skull sex identification has more research value. In order to identify the gender of the damaged skull and narrow the search scope of criminal cases, this paper proposed an incomplete skull sex discrimination model. First, it divided the skull into seven zones, marked the feature points and quantify the unmeasurable features. Then, it selected the optimal feature subset of each partition by using forward stepwise regression method based on maximum likelihood estimation. It set up and tested seven partition sex determination decision models by using leave-one-out cross validation. Finally, it constructed the final sex determination for incomplete female and male skull. The experimental results show that all the subdivisions can determine the sex ofthe skull. The more subarray combinations, the higher the discrimination accuracy.