Research On Fuzzy Logic And Neural Networks Based Medical Images Fusion

The medical image fusion algorithms based on fuzzy logic and fuzzy neuralnetworks is proposed in the paper.Firstly, a new method for multimodal medical image fusion based on T-S fuzzyinference system is put forward in the paper. According to the equivalence principle ofT-S fuzzy inference system and RBF neural networks, the parameters of the T-S fuzzyinference system can be obtained through the training of the RBF neural network. Theinput of the T-S fuzzy inference system is CT images and MRI images, and the output isthe fusion image.Secondly, an image fusion method based on RBF fuzzy neural networks is proposedin the paper. It simplifies the network structure and training algorithms, and improvesthe performance of the fusion. Finally, the area variance is introduced in this paper, andcombined with the algorithms based on fuzzy neural network, the fusion images qualityis improved significantly.Simulation experiments show that the proposed method is feasible and effective.Compared with the traditional methods, the proposed methods have a good performance,especially in dealing with the noise pollution images.