A chaotic neural network-based algorithm for relational structure matching

The matching of relational structures is a problem of prominent importance in pattern recognition research. Since this problem can be transformed into the equivalent problem of finding the maximum clique, the largest sub-graph whose vertices are mutually connected, in a derived association graph, all algorithms, which can solve the maximum clique problem well, will inevitably solve the relational structure matching problem effectively. In this paper, an algorithm based on a chaotic neural network is proposed to solve the relational structure matching problem by finding the maximum clique in the association graph. From detailed analysis, we draw the conclusion that, unlike the conventional Hopfield neural network, the chaotic neural network can avoid getting stuck in local minima and thus yield excellent solution in finding the maximum clique in a given graph. Experimental results also verify that this algorithm is more effective than the conventional Hopfield neural network-based algorithm in solving the relational structure matching problem and thus has a profound application potential in pattern recognition.