Jose J. Bouza

Thanks @cookpa, I was using RAS coordinates. I mapped my coordinates to LPS and checked with itk-snap that they were sensible for my images. I am still seeing the same...

Here is a minimal example of the issue: https://drive.google.com/drive/folders/1aMS0gJFQsDZU9AKWKaTnrpxaLhqffa6z?usp=sharing Running `antsRegistration -d 3 -o "[warped,warped.nii]" --metric "ICP[fixed.txt,moving.txt,1.0]" --metric "MI[fixed.nii,moving.nii.gz,1.0]" --transform "SyN[0.1,0,0]" --shrink-factors 4x2x1 --smoothing-sigmas 3x1x0 --convergence "[1000x100x100,1e-4,4]" -v` Gives: ```...

@ntustison Yes, I did. Correct me if I'm wrong, but I don't think that repository has an example of registering point sets specified as space delimited .txt files. E.g. `antsRegistrationCommandDiffeomorphismTest.sh`...

> @ntustison is there a minimum number of points for the metric to work effectively? The files above only have two points This is a minimal example, I tried with...

> You're right but you can convert those labeled images to .csv point set files using `ImageMath ... LabelStats ...`, ensure that it works as expected, and then map it...

Thanks @ntustison, I can definitely offer some more details (and I appreciate the continued help, I know this is a side-project). I have dug into this for a couple of...

Also relevant: The same problem was touched on in [this thread](https://github.com/ANTsX/ANTs/issues/1222) but never resolved.

By "landmark disks" I just mean that e.g. `data/pointsLowerLeft.nii.gz` in the chicken repository represents a "landmark" as a disk shape region of pixels with non-zero radius. I agree that internally...

Thanks @cookpa. So I think we can safely assume `--metric ICP[fixed.txt,moving.txt,1.0]` does not work because there is insufficient information to define the physical space. So then we have two other...

The mask comment helps a lot. The following command fixes the issue with the segfault from [here](https://github.com/ANTsX/ANTs/issues/1335#issuecomment-1095496896): `antsRegistration -d 2 -o "warped" --masks "[data/pointsLowerLeft.nii.gz,data/pointsUpperRight.nii.gz]" --metric "ICP[data/pointsLowerLeftFinal.txt,data,pointsUpperRightFinal.txt,1.0]" --transform "SyN[0.1,0,0]" --shrink-factors 4x2x1...