A novel optimized rigid image registration of brain using ACMI

The Mutual Information (MI) is considered to be one of the best metric for mono and multi modal registration results in mis-registration due to spatial information lacking. This problem can be solved by using Gradient Coded Mutual Information (GCMI), contains less information than MI. In this paper the advantages of above two methods can be combined, and known as Adaptive Combination of MI and GCMI (ACMI). It also laid stress on optimization step, where the process is optimized using Downhill-simplex method and Gauss-Newton method and results show that the accuracy, robustness, reliability and reduction in computation time are improved with ACMI using Gauss-Newton method.