The Key Technology Of 3 D Face Modeling Multiple Points Of View

Because of the human face contains extremely rich internal and external informations, the study of human face not only is the need of an art remodeling, but also has a wide range of practical value in the field of industrial design, film and television, medical simulation and virtual game, etc. With the rise of computer technology, how to use the computer reconstruction of a realistic3D face modeling has gradually become a hot topic of computer graphics research. Since Parke generated a face model using computer first time in1972, realistic3D face modeling has achieved many fruitful, but due to the complexity of the face itself, the current study has still many problems in the efficiency of the algorithm, realistic effects and a wide range of application.This article comprehensively analyzes the advantages and shortcomings of previous research results, and on this basis it detailed designs and improvements the realistic3D face modeling process, the specific methods and key technologies, uses an approach to reconstruct3D face model recovery from multi-angle face photos. The specific research works include the following aspects:1. Simplified calibration technique of feature key-points. Due to traditional facial feature points manual calibration method is too cumbersome and not accuracy enough, this article designed a simplify feature point editor, by semiautomatic human-computer interaction; it combined the automatic detection and precise adjustment ways. First, using improved ASM algorithm extracted68feature points, then use the space projection control method to mark the corresponding feature points on the model, then apply to the editor manual refinement adjustment, finally, the program will save the group data using this article’s simplified MPEG-4standard automatically.2. Specific human face model mesh generation technology. Comprehensive the advantages and disadvantages of common face model construction method; this article used a method to reconstruct3D face model based on multi-angle face photo, we used the Free Form Deformation face geometry grid as standard face model, then by coarse transformation to build uniform coordinate system, by scattered data interpolation to adjustment non-feature points, by grid refinement process to ensure accuracy and smoothness, in order to achieve the corresponding relationship between the key points, deform a standard model to a specific personalized model. 3. Human face texture panoramic image synthesis technology. This article takes a summary to the commonly used image Mosaic and fusion, and lays special stress on analyzing the simple face texture panoramic image with a weighting factor and the basic principles of image pyramid. On the basis of analyzed and discussed, this article uses an integration of the simple face texture panoramic image with a weighting factor methods and image pyramid with improved Gauss-Laplace operator, mosaic the image in the different pyramid decomposition, then let the each splice layer anti-decomposition to reconstitute a human face texture panoramic image.4. Realistic face model of texture mapping and rendering techniques. Taking into account the3D face can be approximate by the cylindrical surface surrounded, this article uses the cylindrical texture mapping method to establish correspondence between two-dimensional textures to the3D model, and uses OpenGL texture mapping and rendering programming techniques to complete realistic vision of a3D face model.Experiments show that the series of improvement and optimization measures apply in the multi-view3D face modeling process has achieved good results in this article; it obtains double increase of the efficiency and quality in the actual modeling process, and has generally application value.