Texas’ South by Southwest (SXSW) festival is in full swing, with cutting-edge technology on show at SXSW Interactive providing a playground for innovative thinkers with an eye on the future.
China is building a facial recognition database to identify citizens in seconds
At Inkstone, we're marking the occasion with a series on the work of Chinese scientists that could transform the world.
In today’s instalment, we look at one of China’s most ambitious and contentious projects: the plan to build a facial recognition database designed to identify any one of the country’s 1.3 billion citizens in seconds.
The goal is for the system to able to match someone’s face to a database of photos with about 90% accuracy. The project, launched by the Ministry of Public Security in 2015, is being developed alongside a security company based in Shanghai.
It is designed to be connected to surveillance camera networks and will use cloud facilities to connect with data storage and processing centers distributed across the country.
The scale of the national project is unprecedented, and some experts have questioned how long it will take to complete.
Facial recognition is widely used in China – for everything from students’ dorm access to payment for Kentucky Fried Chicken via payment apps.
The government also uses it as a means of maintaining control – it has been used to scan crowds for criminal suspects at events and even to identify and fine jaywalkers – and there are signs that the national database will be used for similar purposes.
Experts have raised some technical questions: Fan Ying, a researcher at the ministry’s population management research center in Beijing, says the project team has encountered “unprecedented challenges.”
When a photo, gender and age range are entered, the system needs to find a match within 3 seconds at an accuracy rate higher than 88%.
When Fan and colleagues tested a facial recognition algorithm developed by Tsinghua University, a world-leading institute in this field, they were disappointed with the results.
The current accuracy rate was below 60%, they said in a paper published in the domestic journal Electronic Science and Technology in May.
“It cannot solve problems with real-life applications,” they added.
A researcher at the Institute of Computing Technology in Beijing familiar with the project warned that some huge technical hurdles remained, saying: “Among 1.3 billion people, some totally unrelated people have faces so alike even their parents cannot tell them apart.”
The researcher also warned that as the system grew, there was a greater risk of data leakage and security breaches.