Time efficient and robust 3-D brain image centering and realignment using hybrid genetic algorithm

Describes a time efficient and robust method to center and realign three dimensional brain images obtained by positron emission tomography (PET). This is primordial for most clinical applications. The method is not affected by brain image defects. This robustness is accomplished using a stochastic sign change criterion as an optimization index. This method takes also 75% less (less than 5 minutes) time than the similar method by Minoshima et al. (1992). The calculation time reduction is achieved by combination of the use of a hybrid genetic algorithm for optimization and an "addition-only" rotation transformation.