I know you're angry that your train is late but try to put a brave face on it - or risk getting nabbed by anti-terrorist police.
Canadian researchers have come up with a new type of facial recognition technology that can detect and classify human expressions.
Dr Prabir Bhattacharya at Concordia University in Montreal is developing a system is to take and analyse photos of individuals in busy areas where security is a primary concern, such as an airport or train station.
Facing up to the truth
Bhattacharya's system measures 15 key points on the face and then compares these against images of identifiable facial expressions. Facial expressions do not actually involve the entire face, but rather specific sets of muscles near the eyes, nose and mouth.
Although there is wide variety in expression across individuals and cultures, the scientist has identified seven basic expressions that seem to be relatively universal.
The next step is processing images fast enough to identify those individuals who might be problematic, whether that's merely being drunk, causing trouble with security staff or even being terrorists.
So the next time you feel the urge to throttle a commuter with tinny earphones or an unhelpful ticket collector, remember to keep a smile on your face...
Your comments (1) Click to add a new comment
mpeskett
December 3rd 2008
1. The number of false positives this kind of monitoring would generate, if used in any busy transport hub, would make it close to useless. Maybe even worse than useless if security is too busy looking at mistakenly flagged stressed out passengers to notice a particularly calm terrorist stroll through.
Looking at the numbers... if they used this at Heathrow (admittedly a very busy airport, but as an example) even if it was able to say with 99.9% accuracy that a person is not a terrorist, that would mean close to 200 false identifications every day.
I have no statistics for the number of terrorists going through Heathrow, so I couldn't comment on how many people it would correctly identify as dangerous, but they've got plenty of innocent people to try and blend in with there.
Alert a moderator
Tell us what you think
You need to Log in or register to post comments