Closed dgobbi closed 9 years ago
The vtkDICOMCTRectifier satisfies most of the points raised in this issue. It regrids the volume, provides a new matrix, and it can work in reverse.
The points that is misses is that it does not create a mask.
Hi @dgobbi Could you please give some advice how to figure out with point "4 - re-color voxels outside the reconstruction volume as "air" "?
DICOM sets the stored values of these voxels to PixelPaddingValue.
I've thought of two ways to re-color the padding voxels as air:
The first is probably the easiest to do.
Thank you. I am not so experienced with this area, so I don't think I can realize it myself. Do you have some example how to do it? Or may be some article about this theme. Any information will be useful for me.
Thanks
I haven't implemented it, so I don't have an example.
All I have is what you see here. Just an idea.
CT images have two features that can complicate image processing:
The vtkDICOMReader has recently been modified so that the PatientMatrix has a yz shear if the gantry is tilted. So, it would be good to have a filter that follows the reader than can use this shear to re-grid the image as a rectangular volume. There are actually several things that this post-reader filter can do:
Perhaps the filter could also work in the opposite direction, take a rectangular volume and produce a series of slices that mimic a specific gantry tilt. This would be useful for image registration, when a rectangular volume must be written with the geometry of a gantry-titled volume.