Hi everyone,
In every I example found on the net (including the examples, which Simple ITK provides), registration is done without adjusting the spacing before. Why not first adjust the spacing and then do the registration?
In most DICOM and NIFTI files you can easily access this kind of data.
What is the reason for that approach?
Hello @miaomiaomiao,
Why is there a need to āadjust spacingā?
The fundamental concept of an image is that it is a regular grid sampling of physical space with a bounded extent. During registration the intensity values at arbitrary, usually non-grid, points inside the bounded extents are used for computing a similarity between the registered images. Consequentially, there is no need to adjust spacing because we are working in a (approximately) continuous world and not a discrete one.
For additional details, please see SimpleITKās fundamental concepts and registration overview pages or the ITK software guide.
That helps a lot! Thank you very much.