540 A NEW ERROR-MEASURE FOR COMPARISON SHAPES

概要

論文の詳細を見る
Many different algorithms for surface mesh optimization with different results were proposed recently, but without analysis and comparison it is impossible to choose the method best suited for the particular problem. So unlike other papers, we do not present any new algorithm for mesh optimization here, instead we introduce a framework for thorough analysis of surface mesh optimization results and comparison of different approaches. Proposed tool includes two error-metrics evaluating changes in characteristics and features between surfaces represented by original and resulting meshes. We provide mathematical proof that introduced metrics are sensitive to degradation or diffusion of the essential surface features and therefore could be effectively used to determine degree of deviation between input and output meshes of the optimization algorithm. As an example the proposed tool is used to compare three different techniques, reflecting basic ideas of approaches to surface mesh improvement problem.
一般社団法人日本機械学会の論文
2005-09-18