The Beijing subway system is experimenting with facial recognition for passengers to pay their fares, and the technology might be applied to the entire system once the database is accurate enough, according to a report by Beijing Youth Daily on Wednesday.

According to the report, the company that operates the system is running small-scale trials of facial recognition on the airport metro line. The Beijing subway system is likely to adopt the technology once the database proves to be 100 percent accurate, the report said.

Zhou Minfeng, a senior analyst at Beijing-based iResearch Consulting Group, told the Global Times on Wednesday that although the current accuracy rate of the technology has exceeded 99 percent, achieving 100 percent accuracy in facial recognition can be difficult.

"Accuracy represents the system's ability to match the face of a passenger with an existing facial photo in the database," Zhou said. "It's difficult to be 100 percent correct, but currently the rate of above 99 percent is high enough for most applications."

Zhou also noted that for the entire system to adopt facial recognition, facial data for each passenger has to be collected beforehand. But given the large passenger flows on the Beijing subway, the subway company might collaborate with companies already possessing a database such as Alipay.

According to statistics on dir126.com, the average daily passenger flow of all 16 Beijing subway lines exceeded 8.5 million in 2018, with the highest one-day figure reaching 13.4 million. 

"Since it's running a system with such massive data requirements, it is possible that the Beijing subway will share databases with third-party payment platforms so that passengers can simply link their accounts to give access to the facial recognition system," Zhou said.