Modeling holographic recognition system by using number-theory convolution transformation

The problem of computer simulation of holographic recognition systems is considered. An analysis is made of a number theory transform method for image processing where the brightness relief is represented by a matrix of binary numbers. The number theory convolution transformation method saves computer time and storage capacity in this case over fast Fourier transformation techniques since it is the amplitude response that is required at the filter output, and therefore the operations of complex multiplication and addition required by FFT are redundant. The number theory transform also gives a more exact representation of quantized data as it is not limited by the round off and overflow problems of FFT. Software limitations are considered.