|
Image registration aims to obtain the best possible spatial correspondence between misaligned datasets. Landmark-based registration involves matching corresponding landmark points for two images. Previous research revealed that such landmark driven schemes are sensitive to the numbers and locations of landmarks. Major work has been done to compute landmark localization through a certain criterion. We have extracted a collection of landmarks using thousands of iterative simulations on a synthetic MRI brain dataset. We believe that the computed landmarks give the best registration results and can serve as ground-truth. However, the dominant features associated with these landmarks are unknown. It is interesting to learn these important features and they are useful to guild landmark placement when different dataset is used, such as breast/prostate MRI.
|
Sign in
to view more information about this project.
