Research On Parameter Calibration Of PCNN Based On Cultural Algorithm

Artificial neural network is one of the most important areas in the Artificial intelligence field. Pulse coupled neural network was advanced by the research of the visual cortex of cats' brain, and became the third generation of the new type neural network model. Featured by the characteristics, Pulse coupled neural network has special advantages that not possessed by traditional image processing techniques, and is gradually becoming one of the hotspots in the image processing field.However, in the models of Pulse coupled neural network, many multiple parameters require manual setting, which is a time-consuming work, and makes the model of Pulse coupled neural network applying inconvenient. In order to implement of automatic setting, a novel cultural algorithm is adopted, which simulates the evolution process of the human society. Cultural algorithm has two evolution spaces, which make the ability of global optimization performance better than that of the traditional evolution algorithms. The internal characteristics of Cultural algorithm are analyzed, the effects of two different belief knowledge conducting the Cultural algorithm to the final optimization results are pointed out. The automatic setting of the parameters of Pulse coupled neural network is completed using the automatic global optimization ability of Cultural algorithm. Based on the work above, Pulse coupled neural network image segmentation based on Cultural algorithm is researched and implemented, and expectation effects are acquired.The creditability and correctness of using Cultural algorithm in the automatic setting of the parameters of Pulse coupled neural network are tested and verified by the results of the experiments, which lays a strong foundation for the consequent research of Pulse coupled neural network and Cultural algorithm.