Ever experienced road rage? Someone cuts you off while you’re trying to merge and next thing you know you’re tailgating them like a NASCAR driver at Fontana trying to get a slingshot off the bank. Then they hit the brakes … "screech-crash-bang" … there goes your platinum rating with the insurance company. What if an on-board emotion detection system could tell that you were getting annoyed and intervene? PSA Peugeot Citroen has teamed up with the Swiss Federal Institute of Technology (EPFL) to develop an emotion detection system designed to recognize signs of irritation and fatigue in a driver’s facial expressions.

Technology and software for reading facial expressions has been used widely in applications from treating autism and depression to market research and brand development. Now researchers at EPFL’s Signal Processing 5 Laboratory (LTS5), who specialize in facial detection, monitoring and analysis, are exploring applications for the motor industry.

Irritation has been identified as a key factor in drivers becoming aggressive. Facial expression reading software can identify seven universal emotions: fear, anger, joy, sadness, disgust, surprise, and suspicion. And though everyone displays their irritation behind the wheel differently, lead researchers at LTS5, Hua Gao and Anil Yüce, concentrated on two of these universal emotions, anger and disgust, as these have been accepted as being similar to irritation.

Helping your car to get to know you better

Firstly, the system was given a series of control inputs, still images of the subjects in both an office setting and in real life situations, such as behind the wheel, so that it could learn to read the emotions under review. Next, using an infrared camera mounted behind the steering wheel, video was taken and later processed using facial monitoring algorithms.

"The rapidity with which the comparison between filmed images and thus detection could be carried out depended on the analysis methods used," said Hua Gao of EPFL. "But overall, the system worked well and irritation could be accurately detected in the majority of cases."

Difficulties in detection of irritation were put down to the complexity and variety of expressions shown. Everyone displays irritation slightly differently. Pursed lips and a scowl, or maybe open mouth and eyes bulging.

"When the test failed, it was usually because this state is very variable from individual to individual," Hua Gao continued. "This is where the difficulty will always lie, given the diversity of how we express anger."

The team also tested a fatigue detector designed to measure the percentage of eyelid closure. Future plans for the project also include using voice recognition or lip reading technology, and reading other driver states, such as distraction.

The next step for EPFL is testing the system in real-time using a more advanced facial monitoring algorithm. However, EPFL and PSA Peugeot Citroen have not as yet revealed how this technology would be applied.

Could we, for example, hear a calming voice saying, "I’m not sure you should be doing that, Dave?" or better still, a proximity sensor and speed limiter that kicks in to ensure you can’t tailgate no matter how much you want to, or maybe a friendly wake-up call if you’re starting to doze.

Source: EPFL