A Novel Multi-focus Image Fusion Method Using NSCT and PCNN
Considering multi-focus images from the same scene, a fusion method using pulse-coupled neural network in non-subsampled Contourlet transform domain is proposed. The input images are performed to multi-scale and multi-direction NSCT decomposition, then both the low-pass subband coefficients and the band-pass directional subband coefficients are input into PCNN. The ignition mapping images are obtained via the ignition frequency generated by neuron iteration. With the neighborhood approach degree of ignition frequency, corresponding subband coefficients are selected and the fused result is obtained through inverse NSCT. Experimental analysis demonstrates that the proposed method retains clear regions and feature information of multi-focus images on a greater degree and has better fusion performance than other existing methods.
Keywordsimage fusion multi-focus image non-subsampled Contourlet transform pulse-coupled neural network
Unable to display preview. Download preview PDF.
- 5.Arthur, L., Cunha, D., Zhou, J.: The nonsubsampled contourlet transform: theory, design and application. IEEE. Transactions on Image Processing 10(15), 3089–3101 (2006)Google Scholar