[Insight-users] regional mutual information

[Severino Fernandez]

I just wanted to ask if the regional mutual information of Russakoff
has been implemented by anybody in ITK.
The reason for my interest is that I am registering earth observation
imagery and the failure of Mutual Information is rather high,
depending on the image structure. Of course, in this field image
structure has a wide larger variability than in medicine.
Just using MI from the original data, combined with MI from the edges
extracted from the same image, the success increased significantly (I
partitioned my image into smaller subimages. Combining both results
the number of correctly registered subimages was larger. From this
results a resampling function is computed for the whole image).
Regards