Multi-focus image fusion via region mosaicing

A new algorithm for multi-focus image fusion via region mosaicing is proposed in this paper. The multi-focus image fusion is formulated as a region labeling procedure to obtain a mask. First, pixel-level labeling procedure is carried out using the Laplacian Energy (LE) as the focus measurement. Then a density-based region growing algorithm is proposed to connect labeled pixels into regions and a threshold is used to determine and remove negligible ones, which are relabeled with an iterative propagation algorithm. Finally, an “all-in-focus” image is obtained by region mosaicking on inputs images with respect to the mask. Experiments show that the proposed algorithm can fully preserve the original information of the input images as well as obtain good visual perception quality. Both objective and subjective experiments validate the effectiveness of the algorithm.