Hm, basically just a slight modification of the gesture/process of selecting something for Lens/Image search, since the screen is basically an image of whatever. Not much to get excited about here.
Most of my searches aren't for something I'm looking at but something I think of.