Research On The Method Of3D Face Modeling Based On Morphable Model

In recent years, reconstruction realistic3D face models gradually become theresearch focus. Along with the rapid rise of the science and technology,3D facemodels have been used widely in the field of virtual reality, human-computerinteraction, multimedia communication, information help, teleconferencing and movievirtual characters. Aiming at this issue, this paper studies the existing reconstructiontechnology, reflected in the following:First, optimize the Morphable Model.First of all, using the Morphable Model technology, we discuss the generalprocess of reconstructing3D facial model, namely the Morphable Model acquisitionand the Morphable Model matching stage. Second, in order to reduce calculation ofreconstructing3D face, we need to reduce dimension processing of Morphable Model,so we use the traditional mesh simplification algorithm to obtain the MorphableModel, then we analyze the advantages and disadvantages. Finally, according to theadvantages and disadvantages of traditional mesh simplification algorithm, theexisting mesh simplification algorithm was improved, and the experimental resultswere analyzed.Second, in the Morphable Model of multidimensional expression.Although reducing dimension of the model can improve the efficiency of the3Dface reconstruction, but this is irreversible. When the dimension of Morphable Modelis too much lower, the reconstruction effect distortion seriously. How to reduce theMorphable Model dimension at the same time keeping the connections between eachdimension model, this paper is to solve the problem.Third, optimal the model matching.The traditional optimization algorithm (randomized gradient descent algorithm)at model matching stage can reduce the amount of calculation model matching in acertain extent, but it also have poor stability and robustness shortcomings etc at thesame time. According to the deficiency of the previous algorithm, this paper puts forward a new model matching method based on genetic algorithm, the new algorithmcan improve the stability of the3D face reconstruction and avoid algorithm into thelocal optimal problem.