An Improved Fuzzy Neuron Network Scheme Using the Modified Genetic Algorithm and Structure

In the process of image filtering, it is a difficult problem to get rid of the distorted points and to keep the detail of the image at same time. So we use the fuzzy neuron network (FNN) to fuse the result of filter of the multilevel FIR-median hybrid (MFMHF) and the result of the median filter with a 5*5 window. It is not easy to set the appropriate parameters for the FNN. Thus, an improved FNN based on genetic algorithm (GA) is introduced to solve the problems of setting parameters of the FNN. Simulation results indicate that after the parameters of the FNN being optimized with GA, the performance of the FNN on image fusion is better than the performance of the common FNN