A method is proposed in which handwritten characters are recognized using feature extraction and an associative memory matrix. In feature extraction, simple processes such as shifting and superimposing patterns are executed. A memory matrix is generated with singular value decomposition and by modifying small singular values. The method is optically implemented with two liquid crystal displays. Experimental results for the recognition of 26 handwritten alphabet characters clearly show the effectiveness of the method. (C) 1998 Society of Photo-Optical Instrumentation Engineers. [S0091-3286(98)00206-2].