Intensity-based robust similarity for multimodal image registration

Abstract

This paper proposes a new intensity-based similarity metric that can be used for the registration of multimodal images. It combines the robust estimation with both the forward and inverse transformation to reduce the negative effects of outliers in the images. For this purpose, we firstly employ the multiresolution technique to downsample the original images, then resort to the simulated annealing method to initialize the transformation parameters at the coarsest resolution. Finally the Powell method is utilized to obtain the optimal transformation parameters at each resolution. In our experiments, the new method is compared to other popular similarity measures, on the synthetic data as well as the real data, and the experimental results are encouraging.

Keywords:

• Multimodal registration
• Forward and inverse transform
• Robust estimation
• Multi-resolution technique
• Optimization

Acknowledgements

This work was partially supported by the Hundred Talents Program of the Chinese Academy of Sciences, the Natural Science Foundation of China (Grants 30425004 and 60121302) and the State Commission of Science and Technology of China (Grant 2003CB716104).