This does seem most likely, but the other thing is that this previous image data is from times when no-one was using the camera. So the question is where did the previous image data come from?
My first guess at an explanation would be that the camera is actually a self-contained little image/video engine that's "running" (i.e., capturing images internally) anytime the phone is powered up, and that under certain circumstances tickled by FaceTime, polling the camera engine results in a cached frame getting kicked out of a camera-internal buffer.
Mind you, that's based on absolutely no actual understanding of the hardware, software or APIs involved.
