Multi-focus color image fusionusing block variance voting

For the multi-focus image, a method to distinguish the focal area and the blur pixels is proposed. The variance of the neighbor pixels of each pixel is calculated and different levels of variance with different weights are denoted where a pixel is in the focal area is estimated by the variance of the neighbor pixels´ voting. For the binary matrix, some isolated 0 or 1 to reduce noise is eliminated. Finally, the fusion is done by the pixels of the focal area of two images. Experimental results show that the proposed method is fast and effective.