Research On Cloud Inference Based Medical Images Fusion Methods

The medical image fusion algorithms based on cloud inference are proposed in this paper.A new artificial intelligence method which is called cloud inference is used in this paper, and a medical image fusion algorithm based on cloud inference is proposed. Both the fuzziness and the randomicity are properly dealt with by the cloud model. Cloud model is used here to construct the multiple rules. The rules are designed according to the max algorithms by using two methods. The inputs of the cloud inference are the CT and MRI images, and the output is the fusion image. The experiment of MATLAB shows that the algorithms proposed here is effective and feasible. The original content of CT and MRI images is better preserved and both the amount of information and the detail resolution are obviously improved. Futher more, the result of 9 gray level rules is better than the the result of 3 gray levels rules.Meanwhile, based on the traditon T-S inference, cloud and neural networks medical image fusion method is proposed here. In this method, cloud model is used to replace the membership function to construct the T-S cloud inference system. According to the equivalence of T-S cloud inference system and RBF choud neural networks, the parameter of RBF cloud neural networks can be trained by the error antipropagation algorithm. Then the parameters are used in the cloud inference system to finish the fusion. The inputs of the cloud inference are the CT and MRI images, and the output is the fusion image. The experiment of MATLAB shows that the medical image fusion method based on cloud and neural networks performs better than the tradition T-S fuzzy inference fusion method. Both the amount of information and the detail resolution are improved. And the result is better than the former proposed cloud inference medical image fusion method because of the advantages of training parameters.