Registration of Organ Surface with Intra-operative 3D Ultrasound Image Using Genetic Algorithm

Abstract

AbstractThe goal of this research is to register organ surface with intra-operative 3D ultrasound (US) image. The algorithm can be used to register pre-operative images, such as MR and CT, with intra-operative US image, or to track organ surface in US image sequence. Simulation experiments were performed to demonstrate the registration of prostate surface with 3D US image in a population of 8 patients. Surfaces were segmented from US images and registered with the same images multiple times to show the accuracy and consistency. Surface tracking was demonstrated by registering the surface obtained from one US image with another US image of the same patient. Genetic Algorithm was used as the search engine to find out the 6-degree transform of the surface that best fit the US image gradients. The algorithm is observed accurate and achieved high consistency in all tested images with an average process time of 9.5 seconds.KeywordsGenetic AlgorithmOrgan SurfaceMedical Image RegistrationProstate BoundaryGenetic Algorithm ChromosomeThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.