Dr Adam J Aviv, a visiting professor at Swarthmore College in Pennsylvania, carried out the attacks by using data gathered by an accelerometer on a smartphone.
Typically this sensor logs phone movements in three dimensions, side-to-side, forward-and-back and up-and-down.
According to a news website, the data gathered as the phone is moved is often used in games to steer or guide an onscreen entity such as a car or a ball.
Working with Matt Blaze, Benjamin Sapp and Jonathan Smith from the University of Pennsylvania, Dr Aviv realised that the data gathered by the accelerometer could also be used to work out where someone tapped on a screen when unlocking a gadget with a Pin or pattern.
In controlled tests, data from accelerometers was captured, exported and analysed to see if it matched a bigger "dictionary" of taps and swipes that had been previously gathered, the report said.
In tests, the software developed by the team got more accurate the more guesses it was allowed, it added.
After five guesses it could spot Pins about 43 percent of the time and patterns about 73 percent of the time.
However, said Dr Aviv, these results were produced when Pins and patterns were picked from a 50-strong set of numbers and shapes.
London: Data captured by smartphone sensors can help criminals determine codes used to lock the gadgets, security researchers have warned.
First Published: Wednesday, January 30, 2013, 12:54