Multi-focus image fusion approach based on CNP systems in NSCT domain

Abstract

Coupled neural P (CNP) systems are recently developed distributed and parallel computing models that are abstracted by the mechanisms of coupled and spiking neurons. CNP systems differ from spiking neural P (SNP) systems in two main ways, namely the utilization of three data units, and a coupled firing and dynamic threshold mechanism for neurons. This paper focuses on the application of CNP systems to solve multi-focus image fusion problems, and proposes a novel image fusion approach based on CNP systems. Based on two CNP systems with local topology, a multi-focus image fusion framework in the non-subsampled contourlet transform (NSCT) domain is developed, where the two CNP systems are utilized to control the fusion of low-frequency coefficients in the NSCT domain. The proposed fusion approach is evaluated on an open data set of 19 multi-focus images based on five fusion quality indices, and compared to 11 state-of-the-art fusion approaches. Quantitative and qualitative experimental results demonstrate the advantages of the proposed fusion approach in terms of visual quality and fusion performance.