摘要
[研究内容与方法]由于我国经济迅速发展,人口不断增加,为了对湖北省女大学生体型进行详细划分,借助非接触式三维人体扫描仪获取了200个人体尺寸数据,从中选取27个关于体型的变量特征;通过对人体体质指数、主要身体比例、人体体表角度进行统计分析,选取人体直立状态下侧面形态的角度作为控制指标,在此基础上利用臀凸角、胸凸角、背入角、肩斜角几个角度进行聚类,运用k-means聚类方法将体型分为6大类,然后结合BMI指数总共将体型细分为18类;最终采用Adaboost算法,对人数存在最多的三类体型进行训练,构建神经网络体型识别模型,实现人体体型自动识别,精度值达93.1%,为体型分类提供参考。
Due to the rapid development of China's economy and the increasing population,in order to make a detailed divi⁃sion of female college students in Hubei Province,the data of 200 body sizes were obtained with the help of a non-contact 3D body scanner,and 27 variable features about body shape were selected.Through the statistical analysis of body mass index,ma⁃jor body proportions,and body surface angles,human side morphology was selected as the control index,the main body propor⁃tion and the body surface angle,and choose the angle of body's side shape as control index.On this basis,the hip convex angle,chest convex angle,back angle and shoulder angle were used for clustering.The body shape was divided into 6 categories by kmeans clustering method,and then it was divided into 18 categories combined with BMI index.Finally,the Adaboost algorithm was used to train the three types of body with the largest number of people,and a neural network body shape recognition model is constructed to realize the automatic recognition of human body shape,with an accuracy of 93.1%,so as to providing reference for body type classification.
作者
沙莎
曹瑞琦
田润雨
江学为
陶辉
张俊
邓中民
SHA Sha;CAO Ruiqi;TIAN Runyu;JIANG Xuewei;TAO Hui;ZHANG Jun;DENG Zhongmin(School of Fashion Design,Wuhan Textile University,Wuhan 430073 China;State Key Laboratory of New Textile Materials and Advanced Processing Technology,Wuhan Textile University,Wuhan 430073 China;Wuhan Textile and Apparel Digital Engineering Technology Research Center,Wuhan 430073 China)
出处
《服饰导刊》
2022年第5期15-20,共6页
Fashion Guide
基金
国家自然科学基金项目(61802285)
湖北省教育厅科学计划重点项目(D20201704)
湖北省服装信息化工程技术研究中心开放基金(184084006)
纺织服装福建省高校工程研究中心开放基金(MJFZ18103)
福建省新型功能性纺织纤维及材料重点实验室开放基(FKLTFM1813)
湖北省教育厅科研计划项目(Q20191703)
屈原故里沉浸式虚拟体验建立与文旅产品开发(BXLBX04480)。
关键词
女大学生
体型分类
体表角度
神经网络
female college students
body type classification
body surface angle
Neural Networks
