對稱性感知之局部形體對應

Abstract

我們提出一個全自動的局部形體對應演算法，可以多樣種類的形體間進行語義式對應(semantic correspondence)，即使在複合式物體上，也能有合理的表現。藉由對稱性偵測的幫助，我們可以取得更高品質的對應結果，在拓普結構的小差異與形體上的小特徵上能有比較準確的判斷。利用物體對稱性為基礎對形體做結構性的診斷，藉此我們的演算法比較不受拓普雜訊影響，並且於複合式物體的局部對應有很大的突破。我們利用統計式的方式來推算出局部對應的結果，並利用完整的對應結果做拓普結構與表面特徵的檢驗以推算更合理的對應。

We propose a fully automatic partial correspondence algorithm that allows matching of a wide variety of shapes with semantically similar, which can solve the topological noise causing by composite object. Through the symmetry detection, we obtain the high-quality correspondence on small topological difference or small feature. Use statistical method to synthesize a partial correspondence, but also evaluate the rationality of final correspondence in global structure of shape to avoid topology inconsistency. Our algorithm is insensitive to topology noise, and it takes a step forward on partial shape correspondence between composite objects