A test pattern may be compared with a reference pattern by placing the two side by side in a coherent beam. If the compared patterns are sufficiently similar, regular parallel fringes will modulate the common Fraunhofer diffraction pattern of the compared samples. The regularity of the interference fringes is a measure of the similarity between the compared patterns. The sensitivity of the method is not disturbed by small absolute and relative displacements of test and reference signals. The effect of even small rotations is, however, severe. The effect of random noise is to reduce the contrast of the interference fringes.
H. Weinberger and U. Almi, "Interference Method for Pattern Comparison," Appl. Opt. 10, 2482-2487 (1971)