I don't know if I agree that it's a good test. If I had two images in front of me of random dots, I cannot superimpose them and see a picture of Marilyn Monroe, no matter how long I can look at them. (ah, actually I could go cross-eyed to make them overlap in my field of vision, but I can't do it mentally).
I could imagine a "spot the difference" type of test would be a good one though. The fields of random dots are identical save for one dot, and you have say where it is. Something like that.
Right, the point of the test is that you can't do it unless you can (relatively quickly) memorize what looks like a random picture of noise. With true photographic memory, that should be possible.
I still don't agree - I'm saying that even if I don't have to remember the field of random dots, even if it's right in front of me, I still can't see an image of Marilyn Monroe. Being able to remember it wouldn't help. If being able to remember it wouldn't help me pass the test, then the test will have a high false-negative rate, and is not a good test of memory.
I could imagine a "spot the difference" type of test would be a good one though. The fields of random dots are identical save for one dot, and you have say where it is. Something like that.