SHIFT-INVARIANT AND ROTATION-INVARIANT INTERPATTERN HETEROASSOCIATION(IHA) MODEL

A shift- and rotation-invariant neutral work using an interpattern het eroassociation (IHA) model is illustrated. The shift- and rotation-inv ariant properties are achieved by using a set of binarized-encoded cir cular harmonic expansion (CHE) functions in the Fourier domain as the network training set. Because of the shift-invariant and symmetric pro perties of the modulus of Fourier spectrum, the problem of locating th e center of the CHE functions can be avoided. Computer simulations and experimental demonstrations that demonstrate the shift- and the rotat ion-invariant properties of the proposed IHA neural net are provided.