Interesting question. I don't know the answer but I'm guessing it would have to do with subtle differences between energies of antibonding orbitals, or possibly just a simple steric difference favoring the isomer where the OH and hydrogen are cis to eachother. This source claims simple steric control:
http://chemweb.bham.ac.uk/~coxlr/Teaching/1st_Year/Carbonyl_Group/Carbonyl_pdfs/lecture%207%20(student%20HO).pdf .
Also are you considering the fact that under kinetic control you have to consider the relative energy of the transition states leading to the two isomers, not the stability of the isomers themselves?
I am kind of surprised to learn that stereoisomeric oximes can even be isolated given the fact that imines cannot be. I'd think resonance donation from oxygen would put more single bond character into the C-N bond and lower the rotational barrier. I wonder what I'm missing. Sorry to derail your question!