If there's a baseline I don't see why it should be rocket science - person X's velocity when working from home is Y and when working from the office is Z. Compare Z and Y and there you have it. Sure, it's not apples-to-apples week-to-week/sprint-to-sprint but the averages are pretty consistent.

If you don't have the baseline there are other ways of getting this figured out, not as accurate but good enough typically - instead of comparing someone's office vs. home velocity you go by the average velocity you're after and see if they can hang with that while working from home (and again, I don't see why this would be difficult to figure out).