Title：Reliable Groupwise Geometric Matching for Computer-aided Engineering, Medicine, Forensics, and Robotics  

Content：Geometric matching computes the correspondence between visual data that are spatially or temporally correlated. In this talk, we formulate the computation of geometric matching between two objects/scenes as finding an optimal bijective mapping function and discuss various criteria that should commonly be considered. We then discuss its generalization to finding partial mapping between geometric data. Finally, we discuss our recent work on improving the robustness of partial matching by simultaneously solving inter-object matching among multiple objects. We also demonstrate a set of scientific applications resulted from our developed matching algorithms.