Facial Recognition Now Used to Bust Cyclists, Drivers in Shenzhen

By Adam Robbins, May 9, 2018

Shenzhen’s “electronic police system” has expanded from one intersection to 40, China Daily reports. The possible targets of the AI-powered facial recognition surveillance system also expanded, from pedestrians to drivers and cyclists.

On May 2, the first day of operation, the system detected “58 cases related to jaywalking and 67 cases involving nonmotorized vehicles on roads,” according to the paper. Offenders are matched against the police database and issued text messages and fines, “depending on the severity of the violation.”

The 40 cameras – half “ultra-HD” cameras with seven million pixel resolution, half conventional with two million pixel resolution – can digitally zoom in to recognize human faces in motion, even if those faces are behind a windshield.

In an interview with South China Morning Post, Wang Jun, director of marketing solutions for the Shenzhen-based supplier Intellifusion, explained: “Cameras with seven million pixels of resolution ensure that the image of drivers’ faces behind the [windshield] are good enough for a facial comparison by our system, but only if the drivers’ facial information has been stored in the traffic police database can our system immediately recognize them.”

Reports have noted four types of violations that will attract the facial scanners’ attention: running a red light, failing to observe traffic signals, driving without a license and “using nonmotorised vehicles on roads.”

We’re looking at you, Mobike street cruiser.

There’s no announcement of which 40 intersections are covered by the new surveillance system, but it doesn’t much matter. The Straits Times reports that Shenzhen's traffic police plan “to expand the surveillance network by adding at least 200 sets of facial recognition devices by the end of this year.”

So follow the rules and keep your cycling to the sidewalks: Big Brother is watching.

[Cover image via Newsweek]