Mobile Technology

CarSafe app zooms in on driver safety

CarSafe app zooms in on driver safety
The front camera focuses on signals from the driver, such as head pose, eyes and blinking rate, which provides the data to infer drowsiness and distraction
The front camera focuses on signals from the driver, such as head pose, eyes and blinking rate, which provides the data to infer drowsiness and distraction
View 2 Images
The front camera focuses on signals from the driver, such as head pose, eyes and blinking rate, which provides the data to infer drowsiness and distraction
1/2
The front camera focuses on signals from the driver, such as head pose, eyes and blinking rate, which provides the data to infer drowsiness and distraction
CarSafe takes advantage of the latest smartphone camera technology, which allows the simultaneous processing of dual-camera streams
2/2
CarSafe takes advantage of the latest smartphone camera technology, which allows the simultaneous processing of dual-camera streams

Researchers from the Dartmouth’s Smartphone Sensing Group have one core mission: make smartphones smarter by levering their embedded sensors to create new sensing applications. As part of their pursuit of techno-cognition, the team sensed that smartphones could make driving safer and conceived CarSafe, the first dual-camera app for smartphones.

CarSafe is an app designed to pick up on dangerous driving behavior and conditions. It was designed in partnership with the University of Bologna and Microsoft Research Asia for Android phones. The front camera provides data about the driver. The app crunches that information with learning algorithms to analyze whether the driver is tired or distracted. Meanwhile the back camera tracks environmental conditions. Combining the power of other sensors such as the GPS, accelerometer and gyroscope, the app is designed to alert the driver about potential risks, in and out of the car.

Using camera switching technology, CarSafe schedules the processing of the two different camera classification pipelines at the right time. The front camera focuses on signals from the driver, such as head pose, eyes and blinking rate, which provides the data to infer drowsiness and distraction. The app uses blink detection algorithms to detect periods of micro-sleep, fatigue and drowsiness. If it detects a risky situation, it will alert the driver by displaying a coffee cup icon on the touchscreen accompanied by an audible alert.

At the same time, the back camera pipeline monitors the distance between cars and lane changing conditions, looking for external signs of risky behavior and scenarios. A red sign will flash on the screen if, for instance, the car is too close to the car in front, also accompanied by an audible alert.

The group’s leader, computer scientist Andrew Campbell, said it is possible that the app could pick up on drink driving, but they haven’t yet developed data models for this application.

The researchers are now working to make the app run faster by exploring new multi-core technology and improving efficiency when processing dual camera streams. The plan is to release the application for a free download later in 2012.

The video below shows CarSafe on the road.

Source: Smartphone Sensing Group

1 comment
1 comment
Stefan Padureanu
What if I have to sneeze ? Usually I do like 3 sneezes at a time, and it takes about 1s for each one to "load" ... if the computer sees me with my eyes closed so much won't it start going apeshit ?