A technique for associating an image is described in terms of thenetwork constraint analysis. Pixels and their gray-values are calledunits and labels, respectively, and a set of n pixels called the unitconstraint set T provides the unit-label constraint set R for memorizedimages. Given an incomplete image X which can have occlusion, noise,distortion, etc., R receives screening by X to yield the reduced set R*(R*⊆R), since the elements of X constrain the elements in R. Adepth first search is then applied to the elements of R* to obtainconsistent solutions, if any, which are associated images. The proposedtechnique is free from the interference among memorized images whichother association techniques suffer from. An iterative technique is alsoproposed for speeding up the depth first search. Performance of theproposed association technique is shown by the experiment employing 26alphabetical letters
展开▼