Multi-focus image fusion method using S-PCNN optimized by particle swarm optimization

Abstract

This paper proposed a novel image fusion method based on simplified pulse-coupled neural network (S-PCNN), particle swarm optimization (PSO) and block image processing method. In general, the parameters of S-PCNN are set manually, which is complex and time-consuming and usually causes inconsistence. In this paper, the parameters of S-PCNN are set by PSO algorithm to overcome these shortcomings and improve fusion performance. Firstly, source images are divided into several equidimension sub-blocks, and then, spatial frequency is calculated as the characteristic factor of the sub-block to get the whole source image’s characterization factor matrix (CFM), and by this way the operand can be effectively reduced. Secondly, S-PCNN is used for the analysis of the CFM to get its oscillation frequency graph (OFG). Thirdly, the fused CFM will be got according to the OFG. Finally, the fused image will be reconstructed according to the fused CFM and block rule. In this process, the parameters of S-PCNN are set by PSO algorithm to get the best fusion effect. By CFM and block method, the operand of the proposed method will be effectively reduced. The experiments indicate that the multi-focus image fusion algorithm is more efficient than other traditional image fusion algorithms, and it proves that the automatically parameters setting method is effective as well.