Fitting Technology Based On Personalizede Body Characteristic Curve Dreiven Fashion Distortion

The problem of personality, reality, real-time existing in virtual3D body model is not onlythe key problem of current clothing CAD related research areas, but also the strategic point inbreaking through the bottleneck of clothing electronic commerce. The body models used inpresent tree-dimensional fitting systems are mainly standard models or approximated modeleddeformed from standard models. There’s a certain gap between it and the human model of largenumber and different shapes in reality. However, along with the development of clothinge-commerce, customers’ requirement of personalization for the fitting is higher and higher. So it isnecessary to study a set of fitting system based on real personalized3D body model. Based on this,this paper presents a fitting system using Kinect to scan the human body and reconstructpersonalized3D body model, then extract human characteristic curves to drive the garmentdeformation. The processes of the system includes the following steps: constructing the3D humanbody model and preprocess it; fitting human characteristic curves; then using the curves to drivefashion distortion; at last show the result of dressed person. Body model and garment model arethe most important two parts of the system.As one of the most important parts of the fitting system, body model is also the key factor todrive fashion distortion, so it is very important to deal with human body models. Body models’generation and pre precession are first introduced in the paper, hen in-depth study is paid into thefitting of characteristic human body curves. This paper first introduces the traditional human bodymodeling method, in view of the problem that the human body model obtained by these ways stillhave a gap with real bodies, a new way of fast and easily scanning the body and reconstructpersonalized body model based on Kinect is proposed in this paper. Since the model consists of alarge quantity of point cloud, the key information of the human body are hidden, there must be auseful computer technology to extract the information in need. The definition of the requiredfeature points and lines are introduced in the paper, and the methods of extracting them arepresented in detail. According to structural features of various parts of the body used methods like slicing, convex hull method of point cloud data to extract the human body featureinformation.Mathematical models are constructed according to the feature of characteristic curvesto get fitting curves. Four elliptical with different radii are used to fit the obtained bust lines. Waistand hip lines fitted with hyperelliptic and used B-spline to curve featured body vertical lines.These fitted curves provided well basis for garment deformation.Secondly, this paper intensively studied the other key part of the system--clothing model,including the construction and distortion of the fashion. Clothing shape is affected by lots offactors, including physical factors and geometric factors, and factors differ due to the body partsthe clothes belong to. It makes clothing simulation the most difficult part of the fittingsystem.When it comes to the fashion closed to the body surfaces, its shape is mainly decided bythe shape of the body style. It means that fashion distortion can be driven by characteristic bodycurves. The deformation algorithm is presented in detail in the paper, and fashion penetrationproblem is well solved in a way.Six sets fashions of different styles or colors are adopted in the experiment to be fitted on twodifferent styles body model, using the distortion way proposed in the paper to change the fashionshapes. It verified the feasibility of this method.