Eye See You
A team of Australian scientists has developed a new AI-powered camera system that can detect if you are too drunk to drive.
But the project is not yet quite ready for wide use, with only 75 percent accuracy, according to researchers from Edith Cowan University in Western Australia, who presented the report. camera project at a computer vision conference earlier this year.
The research, marked by TechSpotThere’s still some work to be done in terms of accuracy, but scientists say the technology holds promise for preventing alcohol-related crashes.
“This is not only beneficial for vehicles equipped with driver monitoring systems and eye-tracking technologies, but could also be extended to smartphones, allowing for more effective detection of alcohol intoxication,” said Ensiyeh Keshtkaran, a PhD candidate at Edith Cowan University and lead author of the study. in a statement about the research.
Big Brother
The researchers compiled a dataset of videos showing 60 people in various stages of sobriety and alcohol intoxication. The researchers then trained a machine learning system to detect intoxication in these videos.
It’s a great concept, but the low accuracy is clearly a big problem. Edge cases will probably be a problem too: will it have trouble telling the difference between drunk drivers and people with disabilities, for example?
There is also the delicate issue of privacy. What happens to the data that vehicles collect from cameras that record drivers and passengers? It is not a theoretical question, since cars already collect data huge amounts of data about drivers.
And finally, reliable technology already exists to prevent drunk drivers from getting into accidents: breathalyzers to unlock ignitions in vehicles.
More about AI: Self-driving cars are biased against dark-skinned people and children, scientists say