ABSTRACT
Image registration entails combining information from two or more images. Once images are no longer viewed in isolation, new constraints and requirements arise. Appropriate data must be acquired; various properties of the data that are not normally prominent have consequences for subsequent data processing and the results that can be achieved. Once the data have been acquired they will frequently have to be transferred from the scanning system (MRI, CT, US, etc.) to a separate computer for subsequent processing. This may include data format conversion, coordinate transformation, distortion correction, intensity correction, and so on, in order to achieve consistent input data for the registration process. Finally, it may be necessary to preprocess the data in some way, for example by image segmentation, to provide the registration algorithms with appropriate input information.
