Effective and Efficient Interpolation for Mutual Information based Multimodality Elastic Image Registration

Mutual information (MI) is a popular similarity metric for multimodality image registration purpose. However, it is negatively influenced by artifacts due to interpolation effects. As a result, registration algorithms performance could be affected. In this paper a novel interpolation scheme is presented. It is both effective and efficient. Effective because it limits the presence of local maxima in the mutual information curve, efficient because it is simple to compute being based on simple and optimized distance measures. The method is validated and compared against other techniques both from performance and time complexity persepectives. Differently from other reference works, which perform tests using rigid transformations, method performance was measured using non-rigid parametric transformation. Experimental results show that the proposed scheme represents a good tradeoff between performance and computing requirements, making it suitable for actual elastic image registration purposes.