Possible new source

Getting back to thinking about my thesis topic, I had pretty much settled on using the Epidemiology literature as my primary source when I came across this paper. I hadn't really thought of image analysis because they are basically solving the inverse problem; using correlation to increase clustering, rather than using the clustering to estimate correlation. I think some of the same ideas apply. At any rate, image analysis is a MUCH more active field of study, so even if only the periphery touches on my problem, it might still be a fertile field.