A Novel Multi-focus Image Fusion Method using Pulse Coupled Neural Network in Nonsubsampled Contourlet Transform Domain

摘要

In this paper, a novel image fusion method is proposed which combines nonsubsampled contourlet transform (NSCT) with PCNN. Firstly, it makes use of the NSCT's shift invariance to suppress the pseudo-Gibbs phenomena around singularities, which appears in the DWT. Secondly, the edge feature is used to motive the improved PCNN model, to retain more edge and texture details. Some experiments are performed in images such as clock, pepsi and book images comparing the proposed algorithm with the SML-CT, PCNN-NSCT and SF-NSCT-PCNN methods. The experimental results show that the proposed algorithm can not only extract more important visual information from source images, but also effectively avoid the introduction of artificial information.