Robust Multi-Band Detail Encoding for Triangular Meshes of Arbitrary Connectivity

The flexibility coming along with the simplicity of their base primitive and the support by todays graphics hardware, have made triangular meshes more and more popular for representing complex 3D objects. Due to the complexity of realistic datasets, a considerable amount of work has been spent during the last years to provide means for the modification of a given mesh by intuitive metaphors, i.e. large scale edits under preservation of the detail features. In this paper we demonstrate how a hierarchical structure of a mesh can be derived for arbitrary meshes to enable intuitive modifications without restrictions on the underlying connectivity, known from existing subdivision approaches. We combine mesh reduction algorithms and constrained energy minimization to decompose the given mesh into several frequency bands. Therefore, a new stabilizing technique to encode the geometric difference between the levels will be presented.