Multifocus Image Fusion Algorithm Based On The Contourlet Transform

Adopting the advantages of contourlet such as multiscale, localization, directionality and anisotropy, a multifocus image fusion algorithm based on contourlet transform is developed [1]. Source images are firstly decomposed to domain of the contourlet transform. The image fusion is then implemented in subbands with different scale and direction combining. The fusion rule with maximum absolute value coefficient is used to fuse the high frequency subbands, while the rule dependent on the image's spatial frequencies and contrast are applied to the low frequency subbands. Finally the fused image is obtained through inverse transform. Experimental results show that the fused image using this algorithm has better subjective visual effect, and is better than the one using traditional Laplacian pyramid algorithm and wavelet based algorithm.