Slashdot Mirror
Police 3D-Printed A Murder Victim's Finger To Unlock His Phone (theverge.com)
An anonymous reader quotes a report from The Verge: Police in Michigan have a new tool for unlocking phones: 3D printing. According to a new report from Flash Forward creator Rose Eveleth, law enforcement officers approached professors at the University of Michigan earlier this year to reproduce a murder victim's fingerprint from a prerecorded scan. Once created, the 3D model would be used to create a false fingerprint, which could be used to unlock the phone. Because the investigation is ongoing, details are limited, and it's unclear whether the technique will be successful. Still, it's similar to techniques researchers have used in the past to re-create working fingerprint molds from scanned images, often in coordination with law enforcement. This may be the first confirmed case of police using the technique to unlock a phone in an active investigation. Apple has recently changed the way iOS manages fingerprint logins. You are now required to input an additional passcode if your phone hasn't been touched for eight hours and the passcode hasn't been entered in the past six days.
Capacitive touch (the metal ring around the iPhone's home button) only works if you are alive. That's what turns on the fingerprint reader so its not constantly checking if there is a finger pressed against it. Capacitive touch detects micro-electric currents that flow through your body, and unfortunately, don't work once you are dead.
http://www.urbandictionary.com...
You don't need to 3D print a finger to fabricate a fingerprint, a simple laser printer is enough:
http://www.instructables.com/i...
The original presentation on beating fingerprint sensors with ordinary laser printer printed copies of fingerprinters, laid on gelatin, published in 2002, is available at:
http://web.mit.edu/6.857/OldSt...
It's quite a good presentation, and was verified by MythBusters in 2011.
https://www.youtube.com/watch?...
Mythbusters even demonstrated that simply printing a fingerprint on paper, and _licking the paper_, created a fake fingerprint good enough to defeat most sensors. There's little reason to think that the commercial fingerprint sensors have gotten any better, though I'd welcome a modern retest with modern cell phone and computer keyboard based sensors.
Basically, the "fuzziness" of fingerprint sensors which allows to identify real fingers with real sensors is enough "fuziiness" to allow them to be beaten with even casually made fake fingerprints. I've seen no good evidence that the necessarysensor and computational "fuzziness" has ever been worked around with even the most expensive modern sensors: I'd welcome any evidence with honestly done tests showing otherwise.