Classification And Recognition Model Of Young Female Upper Body Shape Based On Curve Features Of Human Body Surface

With the rapid development of today’s clothing industry,people’s demand for the individuation and comfort of clothing products is higher and higher,and the personalized customization of clothing is rising.On the basis of the current national standards,the classification of body shape is the premise to meet the comfort and fitness of clothing.This paper makes a detailed classification of women’s body type A and GB 1335 combined with the national standard of female body type 1335.The 3D human body data were collected and preprocessed by symcad anthropometry system and Geomagic studio.Based on the previous research results of human body surface morphology,seven individual surface angles,including pectoral convex angle,abdominal convex angle,dorsal entry angle,dorsal lateral angle,hip convex angle and shoulder oblique angle,are selected to reflect the ups and downs of human body surface curve.Through Pearson correlation analysis,partial correlation analysis and Rtype clustering,the rationality of body surface angle selection was verified.Using the body surface angle as the classification index,K-means clustering analysis was carried out on 230 human body data collected.By comparing the experimental results and based on the characteristics of body surface curves,the somatotypes are finally divided into three categories.Combined with the classification results,the national standard Y,A,B,C four types of body type were classified into two levels.Through the eight types of body types from Yii to biii,the more detailed characteristics of the body shape of young women under the national standard can be inferred.Finally,through the easydl one-stop deep learning development platform opened by Baidu AI platform,a recognition model that can classify human body shape is created.At the same time,the common cloud API interface is released,and the secondary development of body shape recognition can be carried out.