A self-organising Hopfield network for pattern recognition

Although the analogue Hopfield network has been shown to be a plausible computational mechanism for solving combinatorial optimization problems such as the traveling salesman problem and graph partitioning problem, it has not been effective in solving the object recognition problem by attributed relational graph matching. Recently, we proposed an improved Hopfield model to generate homomorphic mapping. However, in order to generate the desired homomorphic mapping, a number of parameters have to be fine tuned. In this paper, a self-organizing Hopfield network is introduced so that the constraint parameter can be learnt adaptively as the output evolves from the initial state to final state. The Lyapunov indirect method based learning approach is employed to self-organise the network constraint parameter. The proposed self-organizing network is applied to solve circle pattern recognition problem.<>