the whole point of this is using and retaining machine vision, ESPECIALLY when they say they have no software to create and maintain a database of the images. In an industrial setting, using machine vision works as a quality tool because of the traceability of the images to document the quality. At the very least, these postal service images should be traceable to date and time as well as location where the photo was taken. Ergo, investigators need not search billions of photos, but only those in a narrow range - I am extremely cynical that there is NO database. When the postal service guy says:
"the Postal Service does not maintain a massive database of the letter images. The scanning machines at the mail processing centers only keeps images of the letters they scan.
“It’s done by machine, so there’s no central area where any of this information would be,” he said. “It’s extremely expensive to keep pictures of billions of pieces of mail. So there’s no need for us to do that.”
he is dissembling. The very fact the photos are retained allows for present as well as future data mining. There is certainly machine vision capability to scan through millions of images to come up with ALL images which contain a suspect name or address. After all, that's what facial recognition does every day all around the world.
