Multi-Focus Image Fusion Based on NSCT and Focused Area Detection- MATLAB PROJECTS CODE
To overcome the difficulties of subband coefficients selection in multi-scale transform domain based image fusion and solve the problem of block effects suffered by spatial domain based image fusion, this paper presents a novel hybrid multi-focus image fusion method.
Firstly, the source multi-focus images are decomposed using the non-subsampled contourlet transform (NSCT). The low frequency subband coefficients are fused by the Sum-Modified-Laplacian (SML) based local visual contrast, while the high frequency subband coefficients are fused by the local Log-Gabor energy. The initial fused image is subsequently reconstructed based on the inverse NSCT with the fused coefficients. Secondly, after analyzing the similarity between the previous fused image and the source images, the initial focus area detection map is obtained, which is used for achieving the decision map obtained by employing a mathematical morphology post-processing technique.
Lastly, based on the decision map, the final fused image is obtained by selecting the pixels in the focus areas and retaining the pixels in the focus region boundary as their corresponding pixels in the initial fused image. Experimental results demonstrate that the proposed method is better than various existing transform-based fusion methods, including gradient pyramid transform, discrete wavelet transform, NSCT, and a spatial-based method, in terms of both subjective and objective evaluations.