Spot identification in two-dimensional patterns by a least-squares template matching.

Abstract

Abstract The coordinates of the center of measured spots are nonlinearly transformed to get an optimal match between the transformed coordinates and the given coordinates of a reference pattern. The parameters of the transformation are determined by the minimum of a function of squared distances between all spots of the sample and of the reference pattern. The algorithm requires a priori defined correspondences between some pivot points in the sample and the reference and treats their distances differently from the others. The parameters of the transformation are the solutions of a system of nonlinear equations; their numerical values are obtained by iteration.