Application of PCA_RBF Artificial Neural Network in Clustering Male Human Body Shapes

With the digital customization era approaching the apparel industry, human body shape clustering has become more and more important for both enterprises and consumers. This paper examines a new method of male human body shape clustering based on PCA_RBF Artificial Neural Network. Firstly, anthropometric data of 200 males from Hunan, Hubei and Jiangxi province are acquired by 3D human body scanner. Furthermore, the different indexes of body shape clustering which contain most information about body shape are extracted by principal component analysis. Then, the RBF ANN model is built to classify the male human body shape. The male human body shape are divided into 4 categories in our national standard, however, male human body shape can be classified into 7 categories fast and accurately by well trained PCA_RBF ANN. The results reveal that clustering male human body shape by PCA_RBF ANN model is technically feasible.