ABSTRACT:
A novel image fusion framework using PCNN has been proposed in this paper and image fusion has been performed by the firing frequency of neurons for PET and CT images. In general, the result of image fusion is affected by neuron parameters, while the proposed algorithm is independent of neuron parameters and dependent on the image gradient parameter. PCNN is built in high frequency sub bands in order to replicate the biological activity of human inspired systems. In traditional algorithms the linking strength of each neuron is set constant or frequently changed based on the features of each pixel. Hence in the proposed algorithm, the linking strength as well as the linking range is determined by the prominence of corresponding low frequency coefficients. The low frequency coefficients not only reduces the calculation of parameters but also the global features of images have also been flexibly made good. The sequences of multi-scale fusion images were obtained by the parallel PCNN and the fused image can be obtained by inverse DMWT at each process of iteration. Experimental analysis illustrates that PCNN outperforms than DWT and SWT from both visual and quantitative prespective.