作者: T.A. Cass
关键词:
摘要: The author focuses on the central problem of image and model feature matching. In particular he defines a geometrical uncertainty features, devises tractable algorithm for determining all feasible sets correspondences given tolerances. A key insight into matching provided by this work is that search should be guided analysis (feasible) transformation space rather than correspondences. This because only interested in maximal match-sets not exponentially many subsets them as found correspondence search. >