Recognition Of Facial Feature And Construction Of 3D Facial Model In Vitual Fitting System

With the rapid development of the internet technology,the electronic commerce has acquired unprecedented development.In order to promote the online clothing sales,research work on virtual fitting technology has become a hot spot.In order to let the consumers agree with that the three-dimensional human body model in the virtual fitting system can represent themselves,the model should accurately emulate not only the human body but also the customer’s face,skin color,temperament and other characteristics.Therefore building the three-dimensional face model plays an important role in the three-dimensional human body modeling.The costomers of online shopping are distributed among the world,so that the three-dimensional model must be universality,simple to operate etc.Based on the single input positive photo,a realistic three-dimensional face model is created.This research work has prepared for the combination model,also laid foundation for further study in Virtual fitting technology.Firstly,based on the skin color has special clustering in YCb Cr color space,the skin gaussian model is setted up.Then identified the face region in the input picture.According to the vertices of Candidie-3 model,we must choose the facial feature points.Secondly,based on some prior facial knowledge and the unique feature of organs and outline,the facial feature points can be traced.Then,based on the identified feature points,we make a series modification on the vertexes of general Candide-3 face model to generate specific three-dimensional face mode.At last,we can build the realistic three-dimensional face model by using texture mapping technology.The results show that the created model has many advantages such as easy to use,low cost,it also has a strong sense of reality.The model can be widely used.