Multi-focus image fusion based on nonsubsampled contourlet transform and multi-objective optimization

To solve the problems that parameter can not be adjusted adaptively in most multi-focus image fusion algorithms, in this paper we propose a new algorithm based on nonsubsampled contourlet transform (NSCT) and multi-objective Biogeography-based Optimization (MOBBO). Firstly, an improved bandpass directional subband fusion method in NSCT fusion framework is proposed by using a region-based fusion rule with consistency detection. Then, multiple objective fusion evaluation indicators are used as object functions and MOBBO algorithm is carried out by multi-objective optimization search. The final resulting image is obtained through inverse transform. The experimental results show that the proposed algorithm has better objective evaluation and visual result than other exiting image fusion algorithms.