Partial volume and resampling (linear interpolation)

Hi there

Sorry to come again with a non diffusion question, but mrtrix has alos simple and powerfull tools like mrcalc and mrgrid

my question is about the different interpretation of the linear interpolation when comparing for instance itk and mrgrid

I compared a downsampling of a binary mask from 0.25 mm to 1mm and found very different results if I use itk or mrgrid. I first though mrgrid was “right” but given the comment of effigies, it may be just two different point of view …

I already post my question on the nibable forum, so if you could directly comment over there

Many thanks