Fusion of multi-focus images based on three channel non-separable symmetrical wavelet filter banks

In order to solve the problems that the Daubechies series of wavelets have not symmetry and the separable discrete wavelet transforms only emphasize horizontal and vertical directions of the decomposed images, a novel fusion method of multi-focus images based on the three channel nonseparable symmetrical wavelets is proposed. A construction method of the symmetrical filter banks relating to this kind of wavelet is presented. The multi-resolution decompositions of the multi-focus image are performed in nonsubsampled mode using the filter banks constructed. The fusion rule that the value is smaller is selected for the low frequency subband coefficients; while for high frequency subband coefficients, the coefficients that the absolute value is larger are employed. The fused image is achieved through inverse nonseparable wavelet transform. The experiment results show that the proposed method has good visual effect and the fused images have good contrast, rich information, high definition and spatial resolution. The fusion performance outperforms the performance of the fusion method based on DWFT.