Content-Aware Model Resizing Based On Surface Deformation

With the fast development of 3D scanning and related techniques,3D models are now widely used as an emerging type of media in 3D games, computer animation, industrial model design, computer simulation, digital cultural heritage protection and so on. Techniques for processing 3D geometry data have also become a hot research topic in computer graphics, which is called Digital Geometry Processing (DGP). No matter in industry or academic world, much attention has been paid on DGP, and techniques have been greatly developed. However, as the related research continues to develop and the applied range is being enlarged, there are still many challenges in DGP.How to efficiently obtain digital geometry models is thus a practical problem. For those models with the same content, but different scalings, it is trivial to re-acquire from real objects or re-design by artists. A natural idea in such scenario is to reuse existing models to create appropriately scaled models. One way of reshaping existing objects is through resizing.By resizing we mean scaling or stretching the object along some orthogonal directions. The simplest approach to resize a geometric model is to globally scale the mesh along some directions (typical axes of the coordinate system). However, this will lead to unwanted distortions of significant features. Clearly, the larger the scale magnitude, the more the visual distortion. However, this distortion is not distributed uniformly across the model. Certain regions such as feature regions tend to be more sensitive, while other regions are insensitive to the scale regardless of its magnitude. This observation suggests that reasonable resizing should use non-uniform scaling according to the local sensitivity, protecting some parts, while possibly stretching others excessively.In this paper, we propose a novel method for content-aware mesh resizing. We first estimate local differential information at each edge, which is then used to derive the resizing sensitivity. Then, we conceptually attach a spring to each edge. The spring coefficient is related to its local sensitivity. Keeping the mesh connectivity unchanged, the spring system will update vertices to an equilibrium state. The equilibrium state naturally corresponds to the minimizer of a quadratic spring-like energy function.In anthor way, reasonable non-homogeneous scaling results can be obtained by iteratively optimizing a quadratic error function.Our method can decrease visual distortions and maintain the feature of the model.In this work, we propose a novel method for content-aware mesh resizing based on surface deformation. Our method could automatically detect these sensitive regions of the model. Then build a quadratic optimization system. Our method measure per-edge sensitivity to guide non-homogeneous resizing.Compared with the seminal work by Kraevoy et al., our method does not need an auxiliary regular grid, and directly deforms the mesh models according to local sensitivity to geometric scaling. The method is efficient, easy-to-implement and produces reasonable scaling results.