Infant image registration based on improved mutual strict concave function measurement

Based on the analysis of characteristics of mutual strictly concave function measurement, Powell optimization algorithm was improved. It can maintain that the column determinant is not zero in the search direction of each iteration, and with the increasing of iteration, the search direction gradually increase the degree of conjugation. Therefore, infant image registration based on improved strict mutual concave function measure has been proposed. The algorithm has been experimented using actual clinical infant brain image, and compared with other registration algorithm. The experiment results show that the algorithm is applicability with situations such as rotation and translation in the MR image due to infants head twist movement at the time of magnetic resonance imaging, the algorithm gets certain good registration results.