A characteristic-preserving technique for lossless and invertible simpification of large complex three-dimensional Triangulated Irregular Network models

Abstract

This dissertation describes a characteristic-preserving, input-tolerating, lossless and invertible TIN simplification technique, called AQSlim. AQSlim is superior to previously published techniques because it possesses this full set of important characteristics that existing methods can only partially match, and performs comparably or better to existing methods on metrics such as input model size, error, visual fidelity and topological preservation. AQSlim can also be applied as the basis of a lossless compression storage format for TIN models