Study On Cnc Manufacturing-oriented Key 3d Portrait Modeling Technology And Its Application

With the development of CNC machining technology and computer technology, many people are no longer content with all kinds of vitural figures in the computer, and pursuit a kind of portrait products. A kind of customized product about figures have emerged, and become a huge new industry, such as 3D statuary, bobby and embossments. Tratiditionally, the human portraits are hand-made. However, the similarity of between the portrait by tratiditional method and the person mainly depends on the experience and skill of the maker, which make the traditional way spread hardly.The highlight of the customized portrait depends on the modeling of features on the face. Currently, the most important means of reconstructing 3D portrait models is using 2D images, which has been widely used in virtual scenes, such as character animation, film, advertisements and game production. These models are primarily non-geometric models for visual effects or texture models using texture mapping technique, which are unable to adapt to machanical manufacture. Portrait models applied to customized products, which can demonstrate the geometry features of the face, are for CNC programming and machining. This paper mainly studies how to establish the geometrical face model efficiently from two orthogonal images and apply the face model to the portrait industry, providing theoretical methods and techniques support for this new industry. Research contents include the following aspects:The paper studied the interactive extract methods of the feature points from front and side images. Firstly, 15 front 2D feature points and 11 side 2D feature points are selected with reference to FDP parameters in MPEG-4 standard. Then, the feature points are extracted by two steps: moving the feature points set and moving a single point. Lastly, 3D feature points are formed by the front and side 2D points.The paper also studied the deforming method of the 3D average face basing on feature points and weighted displacement. Based on the 3D feature points, the 3D face model is matched to the 2D image. The feature points on the 3D average face are moved to the image points.Then, displacements of other points on the model are calculated by the displacements of the feature points. Lastly, the interactive editing methods of 3D model are introduced.For the feature extraction method and the 3D face deforming method, a prototype system has been developed using VC++ platform and OpenGL graphic library. Finally, the methods above are applied to the portrait modeling and a case of portraits combining CNC maching and traditional method is shown. The resutlts of the case showed that the methods in this paper are more efficient than traditional method.