Medical image fusion using m-PCNN

Medical image fusion plays an important role in clinical applications such as image-guided surgery, image-guided radiotherapy, non-invasive diagnosis, and treatment planning. Pulse coupled neural network (PCNN) is derived from the synchronous neuronal burst phenomena in the cat visual cortex. However, it is very difficult to directly apply original PCNN into the field of image fusion, because its model has some shortcomings. Although a significant amount of research work has been done in developing various medical image algorithms, one disadvantage with the approaches is that they cannot deal with different kinds of medical images. In this instance, we propose a novel multi-channel model – m-PCNN for the first time and apply it to medical image fusion. In the paper, firstly the mathematical model of m-PCNN is described, and then dual-channel model as a special case of m-PCNN is introduced in detail. In order to show that the m-PCNN can deal with multimodal medical images, we used four pairs of medical images with different modalities as our experimental subjects. At the same time, in comparison with other methods (Contrast pyramid, FSD pyramid, Gradient pyramid, Laplacian pyramid, etc.), the performance and relative importance of various methods is investigated using the Mutual Information criteria. Experimental results show our method outperforms other methods, in both visual effect and objective evaluation criteria.