Image Recognition Based On Quantum-Inspired Synergetic Neural Network

The key to the study of the synergetic theory is the macro-characteristics of complex systems, a qualitative change. At the end of 1980s, finding the deeply similarity between the spontaneous pattern formation and pattern recognition, Pprofessor Haken advanced a new theory, in which the patten recognition was applied to the synergetic theory, and constructed the synergetic neural network according to the order parameter evolution equation. The main content of this paper is the research of synergetic neural network model and the prototype vector optimization. The main work of the paper is as follows:(1)Enlightened by the idea of ecological and bionic, an algorithm that applies the method of quantum inspired evolutionary programming to the synergetic neural network image recognition, according to the learning rules of the synergetic neural network. Combining the characteristics of quantum computing algorithm with the capacity of global optimization of evolutionary algorithm, the algorithm has many excellent characteristics. The quantum chromosone can carry multiple state information and create new prototype vectors with the methods of random observation. In this way, the individual diversity is able to be kept better and the precocious phenomena can be avoided. Through using the quantum update algorithm, the image recognition rate is improved. Experimental results show the validity of this new algorithms.(2) The quantum immune clone algorithm is used to study the prototype vectors of the synergetic neural network and apply recognition of the image based on clonal selection theory of artificial immune system and quantum computing theory. Cloning operator has a strong ability of local optimization and the quantum crossover have the exchange ability of information.Then the quantum immune clone algorithm is able to optimize the prototype vectors of image recognition. The results of simulation show the algorithm effectively improves the rate of the image recognition comparing with the existing synergy neural network algorithm.