A Human Mannequin Recommendation System Based On Chinese Standard Sizing System For Garment-Men And Specialized Human Bodies

With the development of the Internet and the change of people’s consumption patterns,more and more people choose to shop onlineThe choice of clothing size is a big problem faced by consumers in online shopping.While the common size in garments design is very unfamiliar to average consumers.They want to buy fit clothes,but can’t provide complete and accurate body sizes.Based on this kind of problem,this paper studied the classification,judgment and recommendation of mannequins based on garments size and specialized human bodies(men with a height of 167.5-172.5cm).The main research contents are as follows:(1)Data processing:The data of 20 sizes parts of 3 81 male humans with height ranging from 167.5cm to 172.5cm were extracted from the human body database for statistical analysis.By means of correlation analysis and factor analysis,combined with the Chinese Standard Sizing System for Garment-Men,four common factors were obtained and refined into height value and girth value.The height values are height,back neck height,waist height,hip height,back neck point to waist;girth values include:weight,chest girth,waist girth,hip girth,neck-base girth,back shoulder width,back width,chest width.Regression analysis of girth value and body weight yielded a univariate linear regression equation between dimensions.(2)Cluster analysis:Extracted the bust girth and weight from the girth values as the classification elements,and combined Chinese Standard Sizing System for Garment-Men and BMI index to classify,generated initial clusters,and performed K-Means cluster analysis.6 body types classified by bust girth value and 9 body types classified by weight value were obtained.K-Means clustering analysis was performed directly on the height values to obtain 3 types of body types classified by height values.(3)Mannequin building:idealized the clustering results,and combined the classification by height value and the classification by bust girth value to build mannequins recommended for men’s tops,with a total of 18(3×6)categories;combined the classification by height value and the classification weight value to build mannequins recommended for men’s pants,with a total of 27(3×9)categories.(4)Body type prediction algorithm:Logistic regression,neural network and CART Tree algorithms were used to compare the judgment accuracy,and integrated algorithm was used to improve accuracy.The results showed that,in the separate algorithm,the Logistic regression algorithm was the best.The accuracy of body shape classification by height value could reach 100%,the accuracy of body shape classification by bust girth value was 97.64%,and the accuracy of body shape classification by weight value was 97.11%.In the integration algorithm,the Logistic regression +CART Tree integration algorithm was more succinctly accurate,the accuracy of body shape classification by bust girth value was 98.16%,and the accuracy of body shape classification by weight value was 97.64%.(5)Mannequin recommendation library establishment:For each body type code obtained by body type classification,a visual Mannequin library for recommending tops and pants was established.The prototype of tops and trousers was improved,and the fitting rules of tops and trousers were established according to the body type classification and Chinese Standard Sizing System for Garment-Men.(6)Through the establishment of a mannequin recommendation system,the visual recommendation from human body size to human mannequin and fitting model was realized by connecting research contents in series.In this study,the classification of specialized mannequin based on Chinese Standard Sizing System for Garment-Men was realized,and the appropriate classification judgment algorithm was selected by combining with data mining technology,and the visualization processing was carried out by combining with virtual fitting technology,and the recommendation system of specialized mannequin was explored and obtained.The system is consumer-friendly,meets the complexity and variability of human body shape,promote the digital transformation and upgrading of the apparel industry,and plays an auxiliary role in the personalized customization of garments.