A parallel recursive algorithm - discrete-time cellular neural network (DTCNN) was proposed. A hole filling template was designed based on comparison principle of ordinary differential equations. AN optoelectronic system in terms of spatial area-coding method and defocused multiple image scheme was constructed for execution. The optical pairwise convolution for the designed scheme in the last step coincides with the simulated results. The proposed DTCNN provides a new approach to the study of Hopfield model network.
