The identity of espionage agents is a closely guarded secret for any country's clandestine service. Look no further than the Scooter-Libby CIA leak on one of it's agents for the political embarrassment and loss of an intelligence asset. That’s not to say that many countries’ counter-intelligence programs don’t know who the spies are; often times there just isn’t any solid proof through which the spies can be captured or forced out through the consulate.

Enter facial recognition software. The technology has been around for quite awhile, but the revolution really occurs when its cheap enough to put everywhere. Most point and shoot cameras now include facial recognition technology to facilitate bad photographers focusing on the wrong subjects. Beyond recognizing a face is matching that face. One major milestone was the usage of facial recognition against every spectator passing through turnstiles at a major sporting event many years ago. But back to expanding its ubiquitous scope, common software is able to match subjects like Apple's iPhoto '09. And there are more projects to include Coca Cola's Facial Profiler which matches faces to find "your clone" in an archive user submitted photos linked to FaceBook.

Jumping back to espionage and counter-intelligence, when these tools are a simple commodity, the power of more advanced recognition and matching against vast databases really comes into play. The ability to mine through image databases looking for targets will make it easier for agencies to profile their targets down to their movements and location (since many forget to cleanse EXIF information before posting). Such information can finally allow organizations to cross-reference a person against other suspected individuals to determine an espionage history, confirmation of activity and even reveal repeat acquaintances for rooting out moles and leaks. Google has already demonstrated a prototype quantum image search algorithm using a new processor from D-Wave founded on Qubits. The Google technique uses a visual recognition learning program for identifying previously unknown objects in a database. Assuming that an extension of this technology to faces against the library of images Google's spiders have snarfed up isn't far fetched. An intelligence agent of the near future may have to be groomed early to avoid having their picture taken at all costs beyond simple daily operations. They may even be subject to operating in single locations only once a computer can quickly recognize their presence in multiple locales in key time periods.

On the lighter side, a failure in facial recognition technology is taken on a funny spin in the following YouTube video where an HP camera fails to detect a black man. I'm not sure what's worse, not being recognized by the computer or having the computer identify mathematically that you're ugly.