Multi-focus image fusion

Multi-focus image fusion is the making of two or more copies of a scene with different focus values into a single focus clear Image. In our study, super-resolution based hybrid technique is implemented for multi-focus Image fusion. Firstly, bicubic interpolation based super-resolution technique is applied to all source Images. Thanks to this, the ınformation of source Images are enhanced. Secondly, Stationary Wavelet transform is implemented to decompose source Images. After this processing, the each source Image is divided into four sub-bands. Then for each sub-band Principal Component Analysis is applied. Eigenvalues and eigenvectors are calculated for each sub-band of source Images. Calculated eigenvectors are sorted and selected eigenvector which has maximum amplitude for each sub-band. Using these values, new fused sub-bands are created. And to reconstruct the new sub-bands, Inverse Stationary Wavelet Transform is implemented. To perform quality objectively, the fused Image is resized using bicubic interpolation method. The proposed method's results have more clarify, more edge information and far from spatial distortions. The performance measuing proves that.