Optimum Multiresolution Fusion Scheme for CT/MR Images Based on Region Feature

The fusion technology of multi-modality biomedical images is important for clinic diagnoses. For structure of bone can only be clearly display on CT, while normal and pathological soft tissue are better visualized by MR. An optimum multiresolution fusion scheme for CT/MR images based on region feature was proposed. The low frequency band of the image multiresolution representation is segmented into important regions, in which the soft tissue information from MR image was emphasized and fused with rule 1. Otherwise the line feature of CT was defined as second important regions and fused with rule 2. Each feature of the regions is used to determine the region's degree of membership in the multiresolution representation, and then to achieve multiresolution representation of the fusion result. Furthermore, the high frequency information was fused with rule 3. The final image fusion result can be obtained by using the inverse multiresolution transform. A new optimum fusion scheme combines information from different modalities based on quality performance for fused images after decomposition. Root mean squared error and edge preservation are used to adjust the fusion rule and achieve the optimum weight. Experiments showed that the proposed image fusion method can have better performance than existing image fusion methods.