微表情特征画像在公安人像识别系统中的应用研究  

Applied Research on Micro-expression Features in Public Security Portrait Recognition System

作　　者：王扶尧 郑坤泉 WANG Fuyao;ZHENG Kunquan(Criminal Investigation Police University of China,Faculty of Surveillance Investigation,Shenyang 110854,China;Patrol Special Police Detachment of Zhangzhou Public Security Bureau,Zhangzhou 363000,China)

机构地区：[1]中国刑事警察学院视频侦查系,辽宁沈阳110854 [2]漳州市公安局巡特警支队,福建漳州363000

出　　处：《中国人民公安大学学报（自然科学版）》2020年第3期94-101,共8页Journal of People’s Public Security University of China(Science and Technology)

摘　　要：有着悠久历史的人像研究发展至今,已经从传统的手绘人像到今天的人像生物识别技术。当这项前沿的生物识别技术让罪犯无处遁形时,传统的刑侦手绘人像技术并没有没落和消失,监控的盲区、损坏或者陈年旧案,还不时在困扰着公安侦查办案人员。根据这一难点问题,将人脸微表情特征融入刑侦画像中,使得画像更加具有识别性、特征性、针对性等特质,让侦查人员及市民更加容易辨识疑犯。在此基础上,模拟实验,根据PCA算法的人像图片识别原理,着重刻画比对人像面部部位特征,找准人像比例关系,并运用PCA算法,在MATLAB软件中对已绘画人像以图像降维矩阵处理分析进行人像比对,为公安机关侦破案件开辟新路径,提供新手段。The facial recognition research with a long history has developed from traditional hand-drawn portraits to nowadays'bio-identification technology.While this cutting-edge technology left criminals with nowhere to hide,the traditional criminal investigation hand-drawn portrait has not declined or vanished due to blind spots and damage of the surveillance or unsolved old cases that have troubled Public Security criminal investigation officers from time to time.Aiming to solve this problem,the facial micro-expression features are integrated into criminal investigation portrait,which makes the portraits more recognizable,characteristic and specific,and makes it easier for criminal investigation officers and citizens to recognize suspects.On this basis,some simulation experiments are carried out in accordance with PCA(Principal Component Analysis)algorithm of facial image recognition with emphasis on part facial features and facial proportion relation.The PCA algorithm is used to compare the painted portraits in the MATLAB software with image dimensionality reduction matrix processing and analysis,thus providing new path and technology for Public Security criminal investigation.

关 键 词：微表情 模拟画像 模糊图像处理 人像识别 PCA算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]