SPARSELY CONNECTED HOPFIELD NETWORKS FOR THE RECOGNITION OF CORRELATED PATTERN SETS

A sparsely connected Hopfield network for the recognition of natural, highly correlated video images is proposed. A general design mechanism for the construction of a local neighbourhood structure using a stati stical analysis of an arbitrary given pattern set is suggested. The du ality between learning and dilution is employed and different learning and dilution schemes am discussed. The practical use and the efficien cy of the model are shown in simulations of a large network (N = 12288 ). We use a set of natural patterns with high inter pattern correlatio ns and a high site correlation within each pattern, in which the corre lations are given and not constructed by special rules as for highly c orrelated random pattern sets. The results obtained are analysed for d ifferent coding types of the binary pattern set.