Abstract:  For problems such as noise, artifacts, and loss of details, a novel medical image fusion algorithm based on intersecting cortical model ( ICM) that optimized by hybrid genetic algorithm (HGA)and the non-subsampled contourlet transform(NSCT)is proposed. In the proposed method, the non-subsampled contourlet transform decomposition is first performed on the source images to obtain their mul.tiscale and multi-direction representations. Then, the low-frequency bands are emerged by weighted sumof eight-neighborhood based modified Laplacian and weighted local energy, which can simultaneouslypreserve the energy and extract the details. The high-frequency bands are fused by ICM which parame.ters are optimized through HGA, that can retain more detailed features from source image pairs. Finallythe NSC'T inverse transform is performed to obtain the fused image. It is proved through experiments thatthis method is better than other comparative methods in both visual effect and objective evaluation inde.xes, and has better fusion effect.

Key words: medical image fusion, HGA, ICM, NSCT