Full-screen mobile phones make facial recognition technology popular among consumers, and the application scenarios of facial recognition in life are becoming wider and wider. From the security check at the airport to the face payment in the mall, I believe many people are no strangers.
Recently, a data science company in the UK, DataSparQ, developed a facial recognition system for bars AI Bar, which automatically identifies bar-age customers ages and orders, and the system is currently in trial operation at a cocktail bar in London, Harrild & Sons 5cc.
According to the news from DataSparQ official website, the facial recognition system AI Bar is based on machine learning technology, which can quickly identify customers near the bar bar and display the queue time of each customer on the big screen above the bar. There is a customer avatar above. The numbers show the order in which they are queued.
ThisThe AI Bar can also identify the age of the customer. If it is less than 25 years old, it will mark the customer’s avatar on the screen to remind the customer to prepare their ID card. The bar staff will check it, which means that the minor has to mix in. The bar is getting harder and harder.
However, the purpose of this facial recognition system is not to identify minors, but to improve the efficiency of the bar order to reverse the decline of the British bar industry.
A survey showed that 14 bars closed down every week in the UK. In the year of 2018 alone, more than 700 bars closed in the UK, the annual supermarket The number of beer sold has exceeded the bar since 2014.
People are less and less fond of drinking at the bar. In addition to price factors, DataSparQ believes that long queue waiting times during peak bar consumption are also important reasons,
According to DataSparQ, a third-party survey of 2000 drinkers in the UK was commissioned, and 75% of respondents said they would choose because of the long queue time.Leave the bar.
So what AI Bar has to do is help the bar improve the problem. After the facial recognition system recognizes the customer’s order, if the customer leaves the system and can still remember the location, this solution may be more expensive than some online restaurants that rely on manual calling. .
Of course, the premise is that the accuracy of the recognition is high enough. DataSparQ indicates that the AI Bar currently has an accuracy of more than 80%. This accuracy is obviously not enough when the traffic is large, but the bartender will train the machine in an error-correcting manner to improve the accuracy.
In addition, DataSparQ is also developing more AI applications for the Parallel Bar. For example, a faceTab can add customers to the bar label and only order specific faces, which will improve the efficiency of multiple rounds of order in the bar. In the future, customers can even place orders based on gestures while queuing.
If you can improve service efficiency through AI Bar, it means that the bar is expected to sell more in the same time.liqueur. DataSparQ means that if 48,000 bars in the UK use AI Bar technology, 36.9 million liters of alcoholic beverages can be adjusted each year.
If these features really mature, it’s definitely cool, but like many facial recognition applications, AI Bar has also raised some media and consumer concerns about privacy issues. A few days ago DataSparQ Respond to the face snapshot and other information is indeed stored locally, but Will be deleted after each session and will not store any customer’s face image for more than 24 hours.
The current AI Bar has been trial run at Harrild & Sons 5cc, but it has only borrowed this venue for testing, but not for the customers who really face the bar. It is still unclear how the technology will actually reach the bar.
DataSparQ plans to officially launch the AI Bar at a bar in London at the end of the year. You may have an impression of some of the previous media cocktails, and a real AI bar is getting closer and closer to us.
The title map is from: DataSparQ