Young Female's Body Shape Classification Based On Wavelet Coefficient And Study Of Pattern Production

Non-contact three-dimensional scanning technology has been widely used to the study of garment production of domestic and overseas.Nowadays,researchers of domestic and overseas have the study of body shape classification from dimension parameters,angles of body surface,front and side profiles,different ages and so on.However,the study of surface curve shapes,especially longitudinal section curves which are closely contacted with clothing structure,has been quantitatively studied which are mainly concentrated in the study of human body local characteristics of the longitudinal section curve,not consideration of the overall curves.Therefore,the overall curves need to be quantitatively studied,which can have better characterization of the human body shape of longitudinal section curve shape.This paper studied the classification of human body shape based on wavelet coefficient of the longitudinal section curves,used wavelet analysis to extract signals as the overall characteristics.Low frequency information which are related to the curves can reflect the difference among the body shapes,which can provide a new method for the current body shape classification.The contents of the research are as follows: 1.The 3D point cloud and body size of 264 young female students were obtained using the [TC] 2 scanner.The Imageware13 was used to extract 3D point cloud data of body shapes for preprocessing.2.Geometrical shape analytical approach was used to extract the point cloud of body shapes.The 3D point cloud data was dealt with cubic spline fitting and wavelet denoising.Finally,we needed to extract the wavelet coefficient of longitudinal section curves.3.Using DB index method to determine the optimal cluster number,the body shapes were divided into four categories using the cluster analysis.We acquired the distribution of each body shape in the classification of national clothing standard,different body shapes were described by "national standard body shape + longitudinal section shape ".4.Using the national standard body of 160 / 84 A and practical prototype pattern as an example,we built the prototype pattern of different sub-body prototype,and described the characteristics of prototype pattern.5.We established the model of cognition of young female torso body shape using the LVQ neural network.The recognition rate of test is relatively high,and the total recognition rate of test is 97.4%.