Fisher判别函数在足迹分析性别中的应用  

Fisher Discriminant Function in Footprint Analysis of Gender

作　　者：王东 张建功 汤澄清[1] WANG Dong;ZHANG Jiangong;TANG Chengqing(Criminal Investigation Police University of China,Shengyang 110854,China)

机构地区：[1]中国刑事警察学院,辽宁沈阳110854

出　　处：《科技创新与生产力》2024年第4期117-120,124,共5页Sci-tech Innovation and Productivity

基　　金：基于足型生物特征的个体识别关键技术研究(2022YCYB35)。

摘　　要：结合脚型性别差异研究,为了选取出足迹上与性别显著相关的特征,以便通过Fisher判别法建立函数模型,本研究测量了共7个与性别显著相关的几何特征,通过T检验和Fisher判别法建立了模型;利用Fisher判别分析建立了典则判别函数和Fisher线性判别函数。实验结果显示,选取的特征对模型影响显著,函数模型预测准确率为92.3%。该判别函数模型可以显著提高足迹分析性别的准确率,可为技术侦查提供帮助。Combining with the study of gender differences in foot shape,in order to select features significantly correlated with gender on footprints and establish a function model through Fisher discriminant method,this study measures a total of 7 geometric features significantly correlated with gender,and establishes the model through T-test and Fisher discriminant method;a canonical discriminant function and a Fisher linear discriminant function are established using Fisher discriminant analysis.The experimental results show that the selected features have a significant impact on the model,and the prediction accuracy of the function model is 92.3%.This discriminant function model can significantly improve the accuracy of gender footprint analysis and provide assistance for technical investigation.

关 键 词：足迹 几何特征 FISHER判别分析 性别预测 

分 类 号：DF794.1[政治法律—诉讼法学]