Most of the time, in my experience, goggles fails to even recognize it as text to translate, because it tries to identify objects, too.