非接触式人体测量系统中人体体型分类与自动判别  

Classification and identification of body type in non-contacted body measurement system

作　　者：李晓久[1] 王玉秀[1] 刘皓[1] 

机构地区：[1]天津工业大学艺术与服装学院,天津300160

出　　处：《天津工业大学学报》2007年第5期33-35,共3页Journal of Tiangong University

摘　　要：采用摄影测量技术进行非接触人体测量时,通过摄取人体正面和侧面图像,可从人体图像中直接获得人体长度、宽度等有关尺寸信息,而人体的围度尺寸则需要通过二维数据转换间接获得.以模糊C均值聚类方法进行人体体型聚类,并根据聚类结果,使用核Fisher判别分析方法进行人体体型的自动判别,结果表明应用判别结果进行围度计算可有效提高计算精度.The size in length and width can be acquired directly from the front and side images in non-contacted body measurement system applying the technology of photography measurement, but the girth size can be acquired only by transforming from the 2D data. The calculating accuracy can be upgraded by using the fuzzy cluster- ing function to classify the body types and using kernel-Fisher discriminant analysis to identify body types automatically. The result shows that calculating accuracy can be improved by using this classification.

关 键 词：聚类分析 核FISHER判别分析 体型判别 

分 类 号：TS941.17[轻工技术与工程—服装设计与工程]