LINEAR ASSOCIATIVE MEMORIES WITH OPTIMAL REJECTION TO COLORED INPUT NOISE

It is known that many linear associative memories are sensitive to inp ut noise, To reduce such noise sensitivity, an approach of designing a n optimal-noise-rejection linear associative matrix is presented in th is paper, An optimally biased, as well as unbiased solution, in the se nse of least-mean-square of recall error for colored input noise, is g iven, The theoretical lower bound of error for biased or unbiased reca ll is derived, A recursive computational procedure for the optimal des ign is proposed, Computer simulation and image recall experiments illu strate a superior noise rejection of the proposed approach.