Improved trinary associative memory for character recognition

Character recognition by a trinary associative memory (TAM) neural network model is investigated. Three different inner product thresholding schemes, namely, zero plus average thresholding, arithmetic mean thresholding, and a novel adaptive thresholding are examined. It was shown by a simulation that the novel threshold prescription scheme with the permanent inhibition scheme enhanced the convergence as well as storage capacity of the inner product associative memory.<
>