I've got a couple more messages on the color copier code rumour , excerpts follow.
Before you go taking any of this as gospel, some of these folks are regular correspondents of mine, but all of them write from personal e-mail accounts. Their credibility is strictly what they've told me. Make your own judgements. Having worked with people who know images and use color copiers a lot I've got trouble believing that the serial number could be embedded in all output, because these folks would be sensitive to any sorts of signal that could be consistently measured, but on a digital system that oculd do image recognition fast enough I might buy it.
Neither of my Xerox correspondents found trace of "Richard Longcoat" mentioned in that note from "The Konformist" included in my previous notes.
We'll start with the sole believer first, someone who prefers not to be named, works for Xerox and was at an internal "Xerox knowledge conference" and stopped by the color scientist group.
They first employ algorithms that search any scan for straight lines. If found the device then checks to see if adjacent areas match known patterns of currency. If they do it checks the next adjacent area and so on until it is has a high degree of certainty. Then it records the incident and prints the copy in blue only. (This prevents currency copying). A service tech will notice the incident next service call.
My correspondent points to US Patent 5638496 "Color image input apparatus having color image identifying function" and US Patent 315098 "Methods and means for embedding machine readable digital data in halftone images", and also cautions "Please make it clear that they all do this Canon, Ricoh, HP, Xerox, anyone who makes and sells digital copiers".
I should point out that US Patent 315098 was filed 5 Dec 1994, I know I had a friend using color copies in her work back in 1990, before even the other patent was filed..
Jay (last name omitted 'cause he admits to something that may or may not be illegal, I'll let him make the call) says:
I remember seeing something about it in the operators manual for the CLC-350 one of my customers leased, but I personally don't buy it.
He chalks it up to disinformation and psywar, and points out that as a test he was able to scan a dollar bill and print out the results on both a Color Laser Copier (brand unspecified) and a Fiery.
Tom from Xerox notes that Xerox would have to pass this information around to all of their OEM suppliers. Xerox has a technology called "Data Glyphs", but he's dubious that a reasonable portion of the image to embed this information in could be determined in the real-time environment of the copier.
wledford points out the difficulty in doing this with strictly optical systems:
While never having been on a design team for the hardware, I work in the large format/color graphics market. We have a number of these high end color copiers by canon, xerox and the like. I can see the need, however, it does not seem functionally possible.
First, copiers as they are commonly understood, you place the subject 'on the glass,' press the button and they are copied.
These days however, these high end systems act also as a printer/scanners. That is, you can put subject on a high end flatbed, scan to file, edit/upgrade image in a number of 'design packages' most notable Adobe products, then, send the image to a hardware RIP, and the image is output by that very same color copier.
My question is, are the codes embedded only in images off the glass? In that case, the codes must be embedded in the glass/lens. If that is the case, you merely have to swop the glass, to throw off the 'system.'
Does all output including the printer function have these 'hidden codes?
If that is the case, then the codes must be embedded on the output, either pretransfer or posttransfer. Transfer occuring when an LED transfers the image to a photoreceptive drum. If it occurs post transfer, the codes must be embedded on the photoreceptive drum. These drums are a semi-consumable, that is, the photo receptive layers, (selenium) wears out. They are replaced. Ican't see a conspiracy where there are numerous drums for each machine manufactured, so that each drum as it is replaced, only ends up in its 'predetermined' machine. Unlikely, we deal in these drums, and the photoreceptor police have never monitored where we put the drums. I will say though, any code placed on the drum will appear on every print, in the exact same location. As we begin to have copy quality problems, we examine the drum, sometimes with a magnifying glass to find the offending nick or scratch, then the drum is 'bufffed' or polished, to remove the stray marks.
If the code is embedded pretransfer, that means it must be done by the led. If that is the case, swap the LED with another, or trade one with your friends, and againg, the system is foiled. I do not see the practicality, feasability of this 'coding.'
P.S. we are a Xerox dealer, and we are not 'in on the secret.'
I would be happy to try and answer any questions you may have, hope the above is clear.
Friday, December 10th, 1999 danlyke@flutterby.com