kausion_420
Active member
BBC News
By Christian Mahne
in Sydney
Imagine being watched from the moment you get behind the wheel.
Every glance is tracked and every blink monitored, with your car warning you before you are going to have an accident.
This is now possible thanks to a system that studies the human face to detect fatigue or distraction and then alerts the driver.
The FaceLab system recently won one of Australia's most prestigious scientific awards, the Eureka prize for innovation in technology.
FaceLab has been developed Seeing Machines, an international team of 20 scientists based at Australian National University in Canberra.
They are experts in human-computer interactivity, face recognition for short.
Fighting fatigue
Since 1996 they have been developing FaceLab, a system that tracks and monitors car drivers by cameras mounted on the dashboard.
Facelab system
Head movements are tracked by the system
FaceLab can tell if you are becoming inattentive to the road by working out where you are looking, how many times a second you are blinking and angle of your head.
The first application of the system is in spotting early signs of driver fatigue.
This is a major problem in a country as large as Australia, where it can take 10 hours or more to drive between main cities.
Globally, fatigue is responsible for some 30% of the 700,000 deaths on the world's roads each year.
When you consider that a driver who has been awake for 20 hours has reactions slower than someone over the legal alcohol limit to drive, the consequences of driving tired could not be more serious.
Early warning
"We use two cameras that give us an offset from your face which gives us 3D depth information," explained Seeing Machines market developer Gavin Longhurst.
A car could detect if there is some abnormality in how the car is being driven or not driven safely
Once the computer has that three-dimensional model of the driver it tracks reference points, such as eyes and mouth, using the cameras on the dashboard.
A PC in the car boot interprets the gaze, blink and head angle information coming from the cameras.
If drivers start to display the characteristic early signs of falling asleep, the system can alert them.
At present FaceLab is in the prototype stage. Every major global car company has bought one of the systems from which to develop their own version.
By Christian Mahne
in Sydney
Imagine being watched from the moment you get behind the wheel.
Every glance is tracked and every blink monitored, with your car warning you before you are going to have an accident.
This is now possible thanks to a system that studies the human face to detect fatigue or distraction and then alerts the driver.
The FaceLab system recently won one of Australia's most prestigious scientific awards, the Eureka prize for innovation in technology.
FaceLab has been developed Seeing Machines, an international team of 20 scientists based at Australian National University in Canberra.
They are experts in human-computer interactivity, face recognition for short.
Fighting fatigue
Since 1996 they have been developing FaceLab, a system that tracks and monitors car drivers by cameras mounted on the dashboard.
Facelab system
Head movements are tracked by the system
FaceLab can tell if you are becoming inattentive to the road by working out where you are looking, how many times a second you are blinking and angle of your head.
The first application of the system is in spotting early signs of driver fatigue.
This is a major problem in a country as large as Australia, where it can take 10 hours or more to drive between main cities.
Globally, fatigue is responsible for some 30% of the 700,000 deaths on the world's roads each year.
When you consider that a driver who has been awake for 20 hours has reactions slower than someone over the legal alcohol limit to drive, the consequences of driving tired could not be more serious.
Early warning
"We use two cameras that give us an offset from your face which gives us 3D depth information," explained Seeing Machines market developer Gavin Longhurst.
A car could detect if there is some abnormality in how the car is being driven or not driven safely
Once the computer has that three-dimensional model of the driver it tracks reference points, such as eyes and mouth, using the cameras on the dashboard.
A PC in the car boot interprets the gaze, blink and head angle information coming from the cameras.
If drivers start to display the characteristic early signs of falling asleep, the system can alert them.
At present FaceLab is in the prototype stage. Every major global car company has bought one of the systems from which to develop their own version.
