A new global optimization strategy for medical image elastic registration
Xingang Liu, Qing Ping Yang, Hou Qing, Na Long · 2012
To improve the robustness of current medical image registration methods, we present a new two-phase global optimization strategy. The strategy is composed of a global optimization phase (rough searching) and a local optimization phase (fine searching). The rough searching is based on some sort of random sampling technique in the feasible domain. According to the sampling theorem, when the sampling frequency is higher than the twice of the maximum frequency of a function, the function can be completely reconstructed from these finite sampling points. The maximum (or minimum) value of the function at these finite sampling points is just approximate global extremum. To obtain the exact global extremum, the fine searching is performed in the small neighborhood of the point corresponding to the approximate global maximum value. The strategy can ensure to find the global optimization solution. A set of medical images were used to test the strategy. The experiments show that the strategy is more robust than other elastic registration algorithms.