GTM算法.ppt

A robust Graph Transformation Matching for non-rigid registration ABSTRACT This paper propose a simple and highly robust point-matching method named Graph Transformation Matching (GTM) – It relying on finding a consensus nearest-neighbour graph emerging from candidate matches. – It iteratively eliminates dubious matches in order to obtain the consensus graph. – It uses the local structural information and it is an efficient method for dealing with high rates of outlying matches. Graph transformation process example Graph Transformation Matching Assuming that two sets of corresponding points and of size (where matches ) are found. Algorithm GTM: (1) (2) (3) (4) when (5) , , and (6) iterate A median K-nearest-neighbour (K-NN) graph K-NN graph is constructed as follows: define a vertex vi for each of the N points ,such that An edge (i,j) exists when: – pj is one of the K closest neighbours of pi – where Find Outlier The selected outlier is the pair Optimisation Fig. 12. An example of the conversion from the (a) median KNN-graph (using K =2) represented as an (c) adjacency matrix, to the new three (b), (d), (e) structures introduced for optimisation. Fig. 12. An example of the conversion from the (a) median KNN-graph (using K =2) represented as an (c) adjacency matrix, to the new three (b), (d), (e) structures introduced for optimisation. Once the current outlier has been found, the re-computation of the graphs can be performed in three steps: (i) remove all occurrences of the outlier from and ; (ii) remove the outlier from and (this is done just by setting the value -1 at cell [outlier,1]); (iii) for each occurrence of the outlier in and (in its first K columns) remove the edge and