Measurements are taken over a period of time. I assume that all the stars looked at resolve to more than a few pixels wide on the CCD. In that case you can visibly see the planet travel across the star in successive frames.

Other methods? In theory if the star's rotating fast enough and you see it side-on, then roughly half the star's light will be slightly blue-shifted (travelling towards you) and half will be red-shifted (away from you). As the planet crosses the star, it will block light coming from one half of the star, then light from the middle, then from the other half.

If there's a dip in the red-shifted part of the spectrum, in time transitioning to a dip in the blue, then you know the planet is moving against the star's rotation. If the dip is blue to red then you know it's following the star's rotation.

And my university lecturer (mentioned in the article) would be very disappointed that I can't remember the actual method of how it's done.