Research On The Smooth Algorithm Of 3D Human Geometric Model For Virtual Fitting

With the rise of personalized clothing customization,the research on human body model as the carrier of clothing has attracted more and more researchers' attention.In addition,the human body model has been widely used in aerospace,military,medical,games,animation,textile and other fields.The existing methods of human body model acquisition have the disadvantages of rough modeling,large amount of data,poor real-time performance,difficulty in control,and inability to rapidly express individualized human characteristics in batches,it can not meet the needs of the virtual fitting system,a modeling method with high modeling speed,good feature retention effect,and can realize personalized human body modeling will be the main research direction.In this paper,based on the research results of domestic and foreign three-dimensional human modeling techniques and grid model smoothing algorithms,a parametric human modelling method is designed to generate a geometric human model.For the parametric human modelling,the geometrical human body model has insufficient smoothness and can not meet the practical application.In this paper,through the in-depth study of various grid model smooth processing algorithms,the advantages and disadvantages of various grid model smooth processing algorithms are summarized through experimental comparison and analysis.In this paper,the Sqrt3 subdivision algorithm has the advantages of fast subdivision and slow increase in vertex.Therefore,this paper applies the algorithm to the geometric human model obtained by parametric human body modeling,and realizes the smoothness of the geometric human model based on the Sqrt3 subdivision algorithm.The geometric mannequin processed by the Sqrt3 subdivision method shrinks a lot on the whole,and inevitably loses many details and features,in this paper,by using Taubin smoothing algorithm,the principle of the contraction degree of the model can be further adjusted by adjusting the smooth positive-negative shrinkage factor.The Sqrt3 subdivision algorithm is improved,and an improved smoothing method of the geometric human body model is realized.The parametric human body modeling method designed in this paper is relative to the current main body modeling methods,which improves the efficiency and flexibility of human modeling,reduces the cost of modeling,and can achieve a large number of personalized human body modeling.In this paper,the experimental results show that the improved smooth mesh algorithm can effectively solve the shortcoming of smooth surface in the parameterized human geometric modeling,and realize simple and efficient.Compared with the original method,the feature preserving effect of the geometric human body model surface is improved.