Design a Learnable Self-feedback Ratio-Memory Cellular Nonlinear Network (SRMCNN) for Associative Memory Applications

Abstract

The self-feedback ratio-memory cellular nonlinear network (SRMCNN) structure with learning capability for associative memory applications is designed and analyzed. The architecture was realized the modified Hebbian learning algorithm in the SRMCNN is proposed. It can learn the exemplar patterns and correctly output the recognized patterns in the associative memory. The weights of B template are generated from the product of the desired output pixel value and the nearest five input neighbouring elements as associative memory for all input exemplar patterns. The learned weights are processed in the ratio with the summation of absolute coefficients on the B template to enhance the feature of pattern. As the results shown that it was learned and recognized 8 incompletely exemplar patterns and output correctly pattern. The structure of the SRMCNN with B template and the modified Hebbian learning algorithm for associative memory are implemented in the 9times9 VLSI circuits for the TSMC 0.25 mum 1P5M CMOS technology. The capability of SRMCNN for the more variant exemplar patterns learning and recognition is greatly improved in the auto-associative memory applications