Invariant optical pattern recognition using calculus descriptors
Shoude Chang, Henri H. Arsenault · Optical Engineering · 1994
We present an integral descriptor whose value is invariant with respect to translation, rotation, and scale changes of an object. The descriptors of objects in the input scene can be presented as correlation peaks by means of optical correlators. A triple-invariant pattern recognition method is proposed based on the descriptor. To improve the discrimination, the use of a descriptor vector is suggested.