Mutual information medical image registration based on improved genetic algorithms

Medical image registration technology is an important research subject formed by multi-disciplinary subjects, and it involves information processing, computer technology, medical imaging, and many other important disciplines, which is an important adjunct for clinical diagnosis, treatment, and preoperative planning. For the defects of standard genetic algorithm for its slow convergence, easy precocious and other aspects, this paper makes some improvement on across calculation process in genetic manipulation. Besides, for the problem of mutual information image registration algorithms for its high complexity of computational and slow registration, this paper proposes an mutual information medical image registration algorithm based on improved genetic algorithm which uses normalized mutual information as similarity measure, uses mutual information to indicate fitness value of each individual, selects individuals of good fitness to copy the next generation at the same time to eliminate the bad one; finds the best registration parameters when mutual information function got the optimal solution by updating.