The University of California at San Diego (UCSD) scientists say the goal was to demonstrate how keys are not really secret, according to a post on the UCSD Website. "Technology trends in computer vision are at a point where we need to consider new risks for physical security systems," said Kai Wang, one of the UC San Diego researchers and a computer science graduate student.
The key duplication system use a key's "bitting code," which is based on the series of around six cuts in most U.S. residential locks. These patterns represent a numeric code that instructs how to open the lock. The system analyzes an image of the key, along with the brand and key type, to build a duplicate key. The researchers demonstrated how the system works using either a cell phone camera image, or even a five-inch telephoto lens shooting a key from the rooftop of a campus building.
"This idea should come as little surprise to locksmiths or lock vendors," said Stefan Savage, a computer science professor from UC San Diego's Jacobs School of Engineering, who headed up the research. "There are experts who have been able to copy keys by hand from high-resolution photographs for some time. However, we argue that the threat has turned a corner -- cheap image sensors have made digital cameras pervasive and basic computer vision techniques can automatically extract a key's information without requiring any expertise."
The UCSD researchers haven't released their software publicly.