Multiresolution morphological fusion of MR and CT images of the human brain

This paper compares two types of multiresolution data structure for data fusion, the data pyramid and the data cylinder. Data pyramids are well established, and in recent years nonlinear operators such as alternating sequential morphological filters have been used in their construction. The essential difference between pyramids and cylinders is that pyramids undergo subsampling at each level. This paper compares the fusion of medical images through pyramids and cylinders and also considers the effects of different sampling ratios