Surface data on dynamic topologies
Master thesis by Andreas Tarandi (KTH Royal Institute of Technology) done at DICE.
We implement, combine, and evaluate the work by H.Schäfer (2012) in ”Multiresolution attributes for tessellated meshes” and C.Yuksel (2010) in ”Mesh Colors” for storing surface data such as colors, normals and displacement directly on the surface for 3D geometry in a topology-independent way.
We provide implementation details missing in the original papers and propose and evaluate possible optimizations to the techniques, with a focus on usage in real-time applications.
The technique of storing data directly on the surface provides advantages over the commonly used texture mapping such as the eliminations of discontinuities and seams. The proposed technique is also useful on dynamic topologies where texture mapping is difficult.
We also evaluate the performance of these techniques in a real time application where they perform resonably well, but not better than traditional techniques, leading to the conclusion that they are only suitable to use in situation where traditional techniques fall short.