Digital Image Correlation issues

Hi everyone. I am currently getting to know the DIC and for its purposes I have been using Ncorr. My question is as follows: when we get the displacement data for our current image, it gives us the displacement value in mm for each point, where the point's position is expressed in pixels, is that wright ? Could I possibly having those positions in pixel scale express them in polar coordinates ? So I'd have two matrices, one with the radius length and one with the angle's value ? Could I then express the length of that radius in mm ? I did so and the wideness of my sample, which I set to be 10mm, from the transformed radius data appeared to be just above 3mm, which is clearly wrong. Can anyone help me with this, please. Thank you.