Geometry compression of tetrahedral meshes using optimized prediction

摘要: In this paper we propose a novel geometry compression technique for volumetric datasets represented as tetrahedral meshes. We focus on commonly used predicting vertex geometries via flipping operation using an extension of the parallelogram rule. demonstrate that efficiency is dependent order in which tetrahedra are traversed and vertices predicted accordingly. formulate problem optimally (traversing and) flippings combinatorial optimization constructing constrained minimum spanning tree. give heuristic solutions show can achieve prediction very close to unconstrained tree unachievable lower bound. also significant improvements our new over state-of-the-art approach, whose traversal does not take into account mesh.