A method for the efficient re-binning and shading based correction of intensity distributions of the images prior to normalized mutual information based registration is presented. Our intensity distribution re-binning method is based on the K-means clustering algorithm as opposed to the generally used equidistant binning method. K-means clustering is a binning method with a variable size for each bin which is adjusted to achieve a natural clustering. Furthermore, a shading correction method is applied to reduce the effect of intensity inhomogeneities in MR images. Registering clinical shading corrected MR images to PET images using our method shows that a significant reduction in computational time without loss of accuracy as compared to the standard equidistant binning based registration is possible. © Springer-Verlag Berlin Heidelberg 2003.
|Title of host publication||Biomedical Image Registration : Second InternationalWorkshop, WBIR 2003, Philadelphia, PA, USA, June 23-24, 2003. Revised Papers|
|Editors||J.C. Gee, J.B.A. Maintz, M.W. Vannier|
|Place of Publication||Berlin|
|Publication status||Published - 2003|
|Name||Lecture Notes in Computer Science|