A novel medical image fusion scheme employing sparse representation and dual PCNN in the NSCT domain

With the advent of emerging technologies in the field of biomedical imaging, multi-modality images which reflect different levels of information about imaging area are available. Medical image fusion deals with the problem of integrating complementary information from different modality images into a composite image which provides precise and specific information for clinical diagnosis and better treatment planning. This paper proposes a novel method for combining multi-modality images in the nonsubsampled contourlet domain (NSCT). A sparse representation based approach is adopted for low frequency band fusion and high frequency subbands are fused based on adaptive dual channel pulse coupled neural network (PCNN). The fused image is then reconstructed by performing inverse NSCT. Subjective and objective evaluations proved the effectiveness of the proposed method.