Research on Somatotype Recognition Method Based on Euclidean Distance

Abstract

In order to replace the body shape of customers in virtual fitting, 829 valid female body size in China were obtained through three-dimensional body scanning. Through factor analysis, several parts that have great influence on the body shape were obtained: body height, bust girth, waist girth, buttock girth. According to Chinese national standard for clothing size, the human body is divided into four body types: Y, A, B and C, by the difference between bust girth and waist girth. On this basis, the body types are subdivided according to the clothing size. Taking the women whose body height is 160 cm (157 ≤ body height < 163 cm) as an example, nine female alternate models are extracted by Euclidean distance. After that, the Euclidean distance is used to carry out intelligent somatotype recognition. The verification experiment shows that the method has better recognition ability and can realize the replacement of human body and intelligent somatotype recognition in virtual fitting.