Researchers develop facial AI system to monitor drivers and combat dangerous driving

Researchers at Edith Cowan University (ECU) in Australia have created an artificial intelligence program capable of identifying drunk, fatigued, and angry drivers simply by analyzing video footage of their faces. Nicknamed "Jack of Many Faces," the single system evaluates subtle facial movements, eye blinking, and overall expressions to simultaneously monitor three major causes of road accidents.

According to the research team, the system can detect blood alcohol concentration with nearly 90% accuracy and drowsiness with 95% accuracy. It is also capable of categorizing drivers into three distinct levels: sober, moderate, or severe. Because high levels of fatigue can physically mimic drunkenness, and anger can trigger dangerous road rage, tracking all three emotional and physical states provides a comprehensive assessment of driver safety.

This algorithm is smart, because it can tell the whether a driver is sleepy, just making a facial expression, or affected by alcohol. By separating these factors, it can better understand the driver's real physical state. — Zulqarnain Gilani, one of the researchers involved in the project.

While traditional breathalyzers and blood tests are highly accurate, they are invasive, require active cooperation, and demand on-field presence to administer. In contrast, this new technology operates passively and continuously in real-time, requiring no physical interaction from the driver.

To ensure the technology works effectively at night, the research team developed a companion model that intelligently merges standard color video with infrared night-vision footage. By combining these two distinct types of video streams, the system can accurately extract critical facial geometry in the dark, improving performance.