GeodesicTraining right now treats all the images as 3D.
In CaPTk, we see if an (itk) image is 2D or 3D by loading it as 3D and then checking if a specific dimension is 1 (I forgot which specific dimension but I can check, I think y-axis)
The thing is I'm not sure if CaPTk does some preprocessing to the images to make sure that this specific dimension is 1 in 2D images or if it's an itk thing.
Regardless there is probably a more elegant way. MITK might have something too, but that would mean that GeodesicTraining will depend on MITK.
GeodesicTraining right now treats all the images as 3D.
In CaPTk, we see if an (itk) image is 2D or 3D by loading it as 3D and then checking if a specific dimension is 1 (I forgot which specific dimension but I can check, I think y-axis)
The thing is I'm not sure if CaPTk does some preprocessing to the images to make sure that this specific dimension is 1 in 2D images or if it's an itk thing.
Regardless there is probably a more elegant way. MITK might have something too, but that would mean that GeodesicTraining will depend on MITK.