Artificial intelligence (AI) algorithms installed in cars can continuously scan the driver's face to detect signs of drunkenness.
 |
| New AI algorithm can help detect drunk drivers. (Source: LinkedIn) |
Researchers say cameras built into cars can monitor drivers for signs of intoxication thanks to a new AI algorithm that can determine the level of intoxication with up to 75% accuracy. The application is expected to help reduce drunk driving accidents.
This new device was recently introduced at the Conference of the Institute of Electrical and Electronics Engineers (IEEE) and the Computer Vision Foundation (CVF) of the United States.
The device uses a single color camera that tracks variables such as gaze direction and head position. It can also combine 3D and infrared footage of the driver's face and rearview video of the driver's posture, event logs, and record driver behavior.
The team used video footage of drivers of different ages, drinking habits, and driving experience. The device can record three levels of intoxication: slightly drunk but still sober, mildly drunk, and severely drunk.
“Our system is capable of detecting the level of intoxication at the beginning of driving, allowing us to prevent the possibility of drivers losing control while driving on the road,” said Ensiyeh Keshtkaran, a PhD student at Edith Cowan University in Australia, who worked on the device.
The device is fully compatible with smart devices so data can be easily transferred to smartphones, she added.
“While efforts are underway to integrate driver alcohol detection systems into future generations of vehicles and autonomous vehicles, the problem of some people driving while intoxicated remains a pressing issue,” she said.
The World Health Organization (WHO) estimates that alcohol is the cause of 20-30% of fatal car accidents worldwide. In Australia, where the project was launched, 30% of fatal accidents involve blood alcohol levels above the legal limit.
Source: https://baoquocte.vn/australia-phat-trien-thiet-bi-moi-tich-hop-ai-giup-phat-hien-lai-xe-khi-say-ruou-278876.html
Comment (0)