In lab experiments I believe one of the ways this is usually handled is by having neutral third parties with no knowledge of the targets, images, etc. come in and do the judging. They are asked to match the target to whichever result they feel best fits. ie. Here is a randomized stack of...