Realistic Three-dimensional Face Modeling Technology Research

To create realistic three-dimensional face in the field of computer graphics model is a very challenging task. With development on virtual reality, video production, game entertainment, video telephony applications, this research is receiving increasing attention. Especially since the 20th century 90's, the fast development of computer vision and computer graphics technology provided three-dimensional face modeling technical support, many scholars began to attempt to create a realistic three-dimensional human face model, and achieved remarkable results.Today, many researchers still continue their exploration of this area. On the one hand, realistic face modeling has an important academic significance, is one of the most difficult problems in the computer graphics, computer vision field. While the problem also includes the physiology, psychology, physics and other fields of inter-disciplinary issues. On the other hand, three-dimensional face model has a very wide range of applications, and along with the development of all aspects of technology, three-dimensional model of human face in all aspects will apply deeper and broader in many areas.Here we present a fast three-dimensional face modeling method of human face model, this method divides human face model into a number of relatively independent functional area of movement, and then use the various functional area-weighted method of free-form deformation DFFD general deformation of face model. As a result of the introduction of functional area, control points are importantly reduced, so greatly reduces the computation which can be very good to meet real-time and more authenticity.And then using a multi-source illumination model for rending, through optimizing the intensity of nature light and each light source to realize estimation of target illumination. Complex actual environment could be simulated approximately by multiple light sources. Using texture mapping method vertically maps the input image to 3D geometric human face model. Texture map is generated in real-time, when mapping the two-dimensional images directly to the three-dimensional geometric model, because of the relationship between the block (such as around the ears), model on the part of the district can not obtain the corresponding texture, therefore the texture will have access to some gaps in the region, in order to fill these gaps in the region, using the basis of known area to fill a blank area of the linear interpolation algorithm. Thus achieve a realistic human face modeling.