Medical Image Fusion using Local IFS-Entropy in NSST Domain by Stimulating PCNN

This paper proposes a new multi-modal image fusion technique for medical application. It is based on the Intuitionistic Fuzzy Set (IFS) and operated on the Non-Subsampled Shearlet Transform domain. After decomposition of two input images, the fusion of low-frequency subbands (LFSs) are performed using max-selection rule. While fusing the high-frequency subbands (HFSs), IFS entropy is generated considering a local neighbor-hood region of the HFSs’ coefficients. These IFS entropies are used as stimulating input to a biologically inspired Pulse Coupled Neural Network. Depending upon the (higher) firing counts of the network, the HFSs are fused. Finally, to yield the fused image, inverse NSST is performed taking the fused coefficients as its input. Superiority of the present technique is evaluated through some numerical comparisons with already existing techniques in the multi-modal medical fusion domain.