I am using SimpleITK library for bias field correction algorithm to preprocess on MR images but this is very time consuming because MR image ussualy has high dimension such as 152,640,640 pixel array. I want to ask that :
Is it correct to apply bias field correction after performing the resampling process first? Is this cause a problem?
While it is logical to do bias field correction on the smaller image (which is probably the resampled one), resampling first and then doing bias correction will lead to slightly lower image quality. Since bias field is assumed to continuous and slow changing, it can be done on a down-scaled image, see an example. So you should first apply bias correction and then resampling, even if you estimate the bias correction on a downscaled version of the original image.