基于主成分分析法的乳房体积回归预测模型  

Study of the PCA Based Regression Forecasting Model for Breast Volume

作　　者：张瑞 刘驰[1] 王文彬 Zhang Rui;Liu Chi;Wang Wenbin(Apparel&Art Design College,Xi’an Polytechnic University,Xi’an,Shaanxi 710048,China)

机构地区：[1]西安工程大学服装与艺术设计学院,陕西西安710048

出　　处：《针织工业》2021年第4期58-61,共4页Knitting Industries

摘　　要：为准确把握女性乳房形态,为其推荐合适的义乳产品及再造乳房方案,以18~25岁未孕健康女性为研究对象,将样本集以胸围差用K-means聚类方法分为4类,提出一种基于主成分分析法的多元回归模型乳房体积预测方法。提取9个乳房特征参数,即乳房横径、乳房纵径、乳深、乳房横弧、乳房纵弧、胸围、下胸围、胸宽及右颈点-乳点距,归纳为3个特征因子作为自变量,即造型因子、围度因子、高度因子,乳房体积作为因变量,得出4个乳房体积回归方程。经误差分析得出4个回归方程能够较好地预测乳房体积。In order to accurately grasp the breast shape of women,recommend suitable breast prosthesis products and breast reconstruction programs,the selected research subjects were non-pregnant healthy women aged form 18 to 25 years,which were divided into 4 categories by breast circumference difference using k-means clustering method,a multiple regression model based on principal component analysis(PCA)was proposed to predict breast volume.Nine breast characteristic parameters were extracted(breast diameter,breast longitudinal diameter,breast depth,breast cross arc,breast longitudinal arc,bust girth,under-bust girth,breast width and the distance between the right-neck point and BP points),which then were summarized into three characteristic factors(modeling factor,perimeter factor and height factor)as independent variables and the breast volume used as the dependent variable.Finally,four regression equations were obtained through error analysis,which can better predict the breast volume.

关 键 词：乳房体积 三维人体扫描 主成分分析法 回归方程 预测模型 

分 类 号：TS941.17[轻工技术与工程—服装设计与工程]