Design And Research Of Mesh Denoising Filter

Mesh denosing is a fundamental problem of 3D geometric processing. The research in this area is not only important to the 3D graphics, but also the field of signal processing. In this paper, we designed three new filters and discussed in detail the characteristics of anisotropic filters. With comparison with previous works, we demonstrate that our filter is fast, volume preserving and error-controlling.Particularly we discussed the relationship between anisotropic filtering and the theory of chaos. We also show and explain some interesting phenomena closely related to chaos. The chaos theory can be used to explain a lot of phenomena that cannot be explained by calculus. In the paper, we also try to explain a interesting phenomenon in mesh filtering: the bifurcation. And based up this, we further analyze the foundation of denoisingFor a kind of filter pair we have designed, we give another kind of explanation: using the theory of belief propagation. Through the analysis of previous filters using this theory, we draw the relationship between filtering, chaos and belief propagation. And we propose the assumption that chaos and belief propagation are closely related.