Generalization of the binary pattern matching in image processing

We would like to show a possible generalization of the binary pattern matching by thresholding in such a way that the matching should be less sensitive for over- and under-exposing. Thematching is executed on the basis of the difference between the grayscale values of the image and the pattern. We introduce a distance function Di,j, and a measuring number ri,j that gives information about the error of the matching. The modified measuring numbers r′i,j make the counting easier and matching should become faster. We give the counting algorithm of the estimation and discuss the transformation of the matching into scale invariant.