Geometric Correspondence for Ensembles of Nonregular Shapes

摘要: An ensemble of biological shapes can be represented and analyzed with a dense set point correspondences. In previous work, optimal placement was determined by optimizing an information theoretic criterion that depends on relative spatial locations different combined pairwise Euclidean distances between nearby points the same shape. These choices have prevented such methods from effectively characterizing complex geometry as thin or highly curved features. This paper extends for automatic shape correspondence taking into account underlying individual shapes. is done replacing distance intrashape particle interactions geodesic distance. A novel numerical techniques fast computations surfaces used to extract these distances. addition, we introduce intershape penalty term incorporates surface normal achieve better correspondences near sharp Finally, demonstrate this new method synthetic datasets.