基于身高体重的青年女性躯干形态分类及识别
来源:用户上传
作者:靳守宁 夏圆平 张贝贝 顾冰菲
摘 要:为了快速便捷地识别青年女性躯干形态,获取了304名年龄在18~25周岁青年女性的躯干形态参数,包括身高、体重及围度相关参数,将女青年躯干形态分为3类(“O胖体”“H匀称体”“X瘦体”),并归纳出每类体型的判别规则;同时建立了基于身高、体重的BP神经网络预测模型,实现了胸围、腰围和臀围的尺寸预测。结果表明:依据3类青年女性躯干形态的分类规则,88%基于身高、体重预测的样本都被正确分类,证明本文中基于BP神经网络预测模型进行体型识别的方法可行,可为个性化样板的生成提供技术参考和理论依据。
关键词:形态分类;BP神经网络;围度预测;体型识别;青年女性
中图分类号:TS941.17
文献标志码:A
文章编号:1009-265X(2022)04-0200-07
Young women's body shape classification and recognition based on height and weight
JIN Shouninga,XIA Yuanpinga,ZHANG Beibeia,GU Bingfeia,b,c
(a.School of Fashion Design & Engineering;
b.Zhejiang Provincial Research Center of Clothing EngineeringTechnology;
c.Key Laboratory of Silk Culture Heritage and Products Design Digital Technology, Ministry of Culture and Tourism,P. R. China, Zhejiang Sci-Tech University, Hangzhou 310018, China)
Abstract: For the purpose of quick and convenient recognition of young women's body shape, the body shape parameters of 304 young women aged 18-25 were obtained, including parameters of height, weight and girth. The body shapes of young women were divided into three categories (i.e., O fat body, H Well-balanced body and X thin body), and the discriminant rules of each type were summarized. Besides, a BP neural network prediction model based on height and weight was established to perform the size prediction of bust, waist and hip circumference. The results showed that according to the classification rules of three body-shape types, 88% of the samples predicted based on height and weight were correctly classified, proving that the method of body type recognition based on BP neural network prediction model in this study is feasible, and it can provide technical reference and theoretical basis for generating personalized patterns.
Key words: shape classification; BP neural network; girth prediction; body type recognition; young women
随着信息技术的发展,服装的生产日益趋于智能化。为了满足客户的个性化和多样化需求,服装企业从大批量、少品种的传统服装加工模式向量身定制的生产模式发展[1]。女体体型较男体更为复杂、女装品种繁多,如何制作更为合体的女装成为了当下的研究热点,其中最关键的部分是对青年女性躯干形态进行研究。1986年,中国标准化研究院第一次对全国人体尺寸信息进行采集,基于此人体信息数据库制订了GB/T 1335-1991《服装号型》[2],此后,中国服装号型国家标准又陆续修订了4次,现行标准为GB/T 1335-2008《服装号型女子》。然而,随着现代生活方式的改变,人们的身体形态也发生了一定程度的变化。
近几年,研究学者利用不同的方法对女性体型进行细分,其中包括:主成分分析法、体表角度分类法、特征指数分类法、侧面形态分类法等。Song等[3]从女性的轮廓形状、臀部角度和人体曲线形状出发,进行得分计算对体型进行分类。孙洁等[4]从人体角度形态差异出发将人体体型分为4类,并构建了基于神经网络集成的体型识别模型,该方法可有效区分人体形态差异。Yoon等[5]提出采用三维空间矢量方向角的方法对人体上侧面体型进行合理的分类,此方法可用于基于尺寸提取的个性化样板生成系统。樊萌丽等[6]采用能够粗略描述人体躯干形态的臀肩宽比以及横矢比例等变量,对人体形态进行大致分类,但此方法无法具体描述人体体型。Lee等[7]使用SNU-BM程序(3D人体测量软件)从腹部肥胖的男体侧面图中识别特征部位,并对他们的体型进行分类。余佳佳等[8]考虑了扁平率、身高腰高之比、胸腰差等衍生变量,采用分层聚类法对人体形态进行分类分析。王婷等[9]根据11项人体参数将青年女性躯干形态分为I、H、S、O 4类并计算了各类体型的覆盖率。
nlc202209070957
转载注明来源:https://www.xzbu.com/1/view-15439265.htm