Interpolation artefacts in mutual information-based image registration

Image registration requires the transformation of one image to another so a s to spatially align the two images, This involves interpolation to estimat e gray values of one of the images at positions other than the grid points. When registering two images that have equal grid distances in one or more dimensions, the grid points can be aligned in those dimensions for certain geometric transformations. Consequently, the number of times interpolation is required to compute the registration measure of two images is dependent on the image transformation, When an entropy-based registration measure, su ch as mutual information, is plotted as a function of the transformation, i t will show sudden changes in value for grid-aligning transformations, Such patterns of local extrema impede the registration optimization process, Mo re importantly, they rule out subvoxel accuracy. In this paper, two frequen tly applied interpolation methods in mutual information-based image registr ation are analyzed, viz. linear interpolation and partial volume interpolat ion, It is shown how the registration function depends on the interpolation method and how a slight resampling of one of the images may drastically im prove the smoothness of this function, (C) 2000 Academic Press.