A Game-Theoretic Approach for Nonrigid Medical Image Registration

After represent the deformation as a field of discrete displacements, nonrigid medical image registration problem can be considered as a maximum a posteriori inference problem in Markov random field, which can be optimized using pairwise Gibbs energy minimization technique. To find a minimum, we model the registration problem as a game among pixels, propose a novel approach for nonrigid medical image registration based on game theory, and design a distributed algorithm. In the game, each pixel selects the displacement which maximizes its payoff; the registration is achieved by an iterative process. The novel approach can register any deformation with high accuracy and few algorithm parameters while retain simplicity and scalability. Experimental results show that the game-theoretic approach for nonrigid medical image registration is feasible and effective.