Morphological Associative Memory Networks And Application Research

Artificial neural network is an important branch of complex network, and morphological associative memory is an important part of it. As we all known, the different of network structure have different physical dynamics characteristics. The research work is focused on the morphological associative memory neural network in this paper, primarily studied the complex morphological bidirectional associative memory and its performance, and application of fuzzy morphological associative memory in pattern recognition. The structure of artificial neural network, learning rule, characteristics of neural node, characteristics and application fields of artificial neural network is introduced in detail in chapter two of this article, and the principles of morphological associative memory is presented.With the deepening of research in the theory and application of artificial network, real neural network becomes powerless in deal with complex signal and complex pattern. Therefore, the artificial neural network in complex field becomes a research hotspot for domestic and foreign scholar. A complex morphological bidirectional associative memory is proposed based on existed complex morphological associative memory by comparing the size of complex with the means of partially ordered relation and changing the topology of the network in chapter three of this paper. And the performance of anti-noise ability, stability and storage of the network is discussed. At last, the simulation experiment has proved the effectiveness of the network.Fuzzy neural network is an organic combination of artificial neural network and fuzzy logic system, which has been widely applied to deal with the uncertainty and fuzziness information of world. The application of pattern recognition by means of fuzzy morphological associative memory was researched in the fourth chapter of this article. And the theory knowledge of fuzzy morphological associative memory network, its relation performance, principle of pattern recognition and several identification methods were also introduced. Using the fuzzy morphological associative memory method can identify the grayscale image with the noise to some extent, and recognition results show that this method have some tolerance for grayscale image that contain certain noise.