Constrained Resizing Of3D Models Based On Combination Method

With the fast progress and development of3D scanning and related techniques,3D digital geometry models have become an emerging digital media type following the sound, image and video, and are widely used in computer animation, industrial model design, digital cultural heritage protection,3D games, movie special effects, computer simulation and so on. Techniques for processing3D geometry data, which are called Digital Geometry Processing (DGP), have also become a hot research topic in computer graphics, and occupy an increasingly important position in the field of computer graphics. DGP includes many aspects, such as geometric modeling, processing and application of the3D digital models. Though over the past decade, the study and use of DGP has attracted extensive attention of academia and industry, and the technology in this direction has also made exciting progress, DGP is relatively young. With the deepening of research and increasing of application scope in the future, DGP is still full of charm and challenge.On the basis of the predecessors, this thesis focuses on how to efficiently obtain digital geometry models. For those models with the same content, but different scaling, it is trivial and inefficient to re-acquire from real objects or re-design by artists. Taking into account that the discrete digital geometric models, such as the3D mesh, generally have great flexibility, a natural idea in such scenario is to reuse existing models to create appropriately scaled models, which is called model reuse. Therefore, aiming at the model reuse, this paper makes a depth discussion around the deformation technology on the3D digital geometry models to seek efficient methods to get new3D models quickly.It is a very common problem in model reuse to resize the existing3D model to adapt to different application scenarios. To resize models means to scale models along orthogonal directions. The simplest approach to resize a model is uniform scaling, which means to scale the mesh globally along some direction (such as axes of the coordinate system). But uniform scaling always can’t meet the demand, because it will destroy some sensitive characteristics and result in unnecessary distortions. We observed that only some parts of the model are more sensitive and easier to distort while scaling the model, which we call the sensitive region. This shows that a better3D model resizing method should add constraint to the scaling process, and be able to achieve non-uniform scale and maintain the characteristics of the sensitive region in the deformation process.In this paper, a new method for resizing3D models while maintaining or scaling specific feature regions flexibly is presented. At first the given model is divided into three different regions interactively, then each region is mapped to a B-spline surface and control points are computed reversely, at last these three regions are resized respectively and combined together optimally. Our method can decrease visual distortions and maintain the feature of the model.This method has no requirement on the topology of given models. The complexity of this method is just irrelevant to the scale, so it is robust and suitable for processing large-scale objects. The method is efficient, easy-to-implement and produces reasonable scaling results.