Now, I have measured the pulse I used to create the pixel.

Capacitor 350V 220 uF was discharged into the coil.

464 A peak current.

Then I measured the size of coil, and unwound it to count the number of turns.

R = 0.5 Ohm
thickness 2.4 mm
inner d 1 mm
outer d 3.2 mm
top dia 2.2 mm
cross section area (approx) 2.0 mm2
N turns = 50

=> average current density in coil = 11600 MA/m^2

FEMM results

On this picture, I've set up color plot to span a range 0.5-3.5T. This range corresponds to magnetic field required to magnetize NdFeB: 0.5 is about where it starts to magnetize, and 3.5 is to achieve saturation.

I've taken these values from http://dx.doi.org/10.3989/cyv.2005.v44.i3.385

This isn't particularly correct, because magnetization affects magnetic field, and that can change what is actually magnetized. But I don't know how to simulate the process properly, so I'll stick to simple treatment for now.