A fast medical image registration using feature points

In feature-based registration approach, the feature matching step and transform model estimation step are associated and usually conducted in a sequential way. In this paper, we propose a fully non supervised methodology dedicated to the fast registration of medical images. We adopt a method that integrates the both steps into one. First, the points of interest are automatically extracted from both images. Then, an algorithm based on Hough transform gives directly a rigid transformation, which allows information transfer between both modalities. Validation study is conducted under a series of controlled experiments and test on IRM images and a comparison study where the matching of the extracted points is established with a correlation technique combined to a relaxation technique. The LMS method is then used to estimate the registration parameters.