Multi-focus Image Fusion Algorithm Based on Regional Firing Characteristic of Pulse Coupled Neural Networks

Multi-focus image fusion aims at overcoming imaging cameras' finite depth of field by combining information from multiple images with the same scene. In this paper, a regional firing intensity (RFI) is defined, which is based on the statistical characteristic in local window of neuron firing times when pulse coupled neural networks (PCNN) is utilized in the image fusion. A novel image fusion algorithm based on regional firing characteristic PCNN (RFC-PCNN) is proposed and RFI is considered as a determination to select the coefficients of source images. First, a multiscale decomposition on each source image is performed by discrete wavelet transform. Second, PCNN is employed to extract features of source images in wavelet domain. Thirdly, RFI is computed and used to combine the coefficients of source images. Finally, the fused coefficients are used to reconstruct the fused image by an inverse discrete wavelet transform. Experimental results show that the proposed algorithm outperforms the wavelet-based and wavelet-PCNN-based fusion algorithms.