Image registration is a very active research area in computer vision, namely it is used to find a transformation between two images taken under different conditions. Point matching is an image registration approach based on searching for the right pairing of points between the two images. Prom this matching, the registration transformation we are searching, can be inferred by means of numerical methods. In this paper, we propose a scatter search (SS) algorithm to solve the matching problem. SS is a hybrid metaheuristic with a good tradeoff between search space diversification and intensification. On the one hand, diversity is basically introduced from a population-based approach where systematic combinations of subsets of solutions are performed. On the other hand, intensification is achieved with a local search procedure, to ensure the local improvement of promising solutions. Our computational experimentation in a real-world inter-subject medical registration environment establishes the effectiveness of our procedure in relation to different approaches usually applied to solve the problem.
展开▼
