by Graham Murdoch: Let´s say Al has a message he wants to send to his friend Bob. He encrypts it with a â€keyâ€-a sequence of 0s and 1s that unlocks the message. In quantum encryption, he sends this key using one of two intertwined alphabets. An eavesdropper has to guess which alphabet to read. A wrong guess, and she´ll be spotted.

Here, our alphabets are shape (a circle means a 0, a triangle a 1) and color (green means 0, orange means 1). Bob reads the shape alphabet-Al´s message of 1, 0, 0 is encoded as â€triangle, circle, circleâ€-and later confirms that Al sent the key as shapes.

But if an eavesdropper tries to read the message, half of the time she´ll read the wrong alphabet (a typical key requires thousands of such guesses). When she does, the quirky rules of quantum mechanics dictate that reading one alphabet randomizes the information in the other. When Bob checks with Al later, he´ll quickly spot the eavesdropper. Photo by Graham Murdoch

PROTECT YOUR KIDS

Cameras with behavior-tracking software will watch for danger on the playground

THE THREAT:
A kidnapper loitering in front of the local school. A thief slipping into a warehouse. A suspicious stranger dropping a bag in a hallway. The average security camera can catch all these actions, but the footage isn´t much help if the person monitoring the screens is dozing, or so wiped out after hours of viewing that he wouldn´t notice a mushroom cloud erupting on monitor three.

THE SOLUTION:
In the next five years, the security industry could be revolutionized by the spread of smart cameras that pick out shady behavior. Like a friend rousing you from a nap on the couch to check out a great play in a ballgame, these cameras will flag important events and encourage security guards-and high-school principals-to take a closer look. Instead of blankly staring at a screen for hours, says David Abrams, the CTO of Interact Public Safety Systems in Winston-Salem, North Carolina, â€you´re only looking at video when the software thinks something is going on.â€

As part of a test program at Hazard High School in Hazard, Kentucky, Interact has set up 16 smart cameras that can pick out suspicious parked cars, intruders climbing over the fence, or people entering buildings at odd times. The software ignores the background elements in the camera´s field of view-buildings, trees, bushes-then establishes a set of normal, baseline events and watches for unusual occurrences. Since it´s all digital, Interact´s system also enables easy Web access. The principal can view suspicious activities on a handheld wireless device, and police responding to a school shooting could tap into a live video feed of the scene from inside their squad cars.

THE ETA:

Although basic test programs are in place now, by about 2012, smart cameras should be capable of much more, says Vaidhi Nathan, CEO of San Joseâ€based IntelliVision. This will stem in part from the spread of higher-resolution cameras, but industry experts also expect the algorithms themselves to become more reliable and versatile. At the University of California at San Diego, computer-vision researcher Sangho Park is developing algorithms that distinguish violent from friendly behavior. Given a camera with high enough resolution, he says, his program can distinguish a friendly handshake from a punch, or a hug from a push, with 80 percent accuracy. Whether it´s behavior tracking or face recognition, though, the basic idea will remain the same. â€The computer does all the work for you,†says Dilip Sarangan, an industry analyst at global consulting firm Frost & Sullivan. â€It makes life a little easier for everyone.†Except the bad guys. -Gregory Mone