Really, if pad_x and pad_y, are negative, they should be passed to the crop filter.
Probably actually the crop and pad filters could be removed and dealt with directly in the resampling filter, but I couldn't figure out with confidence what to set the Origin to.
This doesn't fix the z - my images actually come out with a z size of 108 rather that 127. But I wasn't confident in how to deal with the cropping.
This PR seems to work for my data, visually the umap and NAC PET align.
Partial fix for #14
Really, if pad_x and pad_y, are negative, they should be passed to the crop filter.
Probably actually the crop and pad filters could be removed and dealt with directly in the resampling filter, but I couldn't figure out with confidence what to set the Origin to.
This doesn't fix the z - my images actually come out with a z size of 108 rather that 127. But I wasn't confident in how to deal with the cropping.
This PR seems to work for my data, visually the umap and NAC PET align.