A prototype of the camera is on display at the Roads and Transport Authority (RTA) pavilion at the ongoing Gitex Technology Week in Dubai.
A staff explained to Khaleej Times that two cameras will be installed in the bus - one is a regular CCTV camera to monitor the interior of the bus and the other is the machine-learning (ML) camera with facial recognition.
"The camera with facial recognition will be able to sift out which passengers are not tapping their Nol cards," the staff explained.
"As soon as a passenger boards the bus, the ML camera will map the facial features and see if that person has tapped his/her Nol card. That person may have tapped the Nol card but it has not enough credit - that will be a violation. Or maybe the person has just went past the Nol counter without swiping the card, then the camera will be able to pinpoint that person by using intelligent algorithms that can detect every passenger who boarded the bus," the staff added.
The ML camera is connected to the RTA's Enterprise Command & Control Centre and an inspector will be notified to catch the dodge evader and charge the appropriate fine, which is Dh200.
"The facial recognition technology will reduce the need for human inspectors to conduct one-to-one check of passengers because the camera can easily identify the fare dodgers," according to the RTA staff.
Speaking to Khaleej Times, Shukri Eid, managing director of Cisco, the technology company that partnered with RTA, said: "Fare evaders are resulting in significant lost revenue for the RTA. Using video analytics and the power of AI (artificial intelligence), Cisco can detect fare evasion."
However, the camera with facial recognition is still in its proof-of-concept stage and authorities are still studying its feasibility and verifying its practical use. Once rolled out, the facial recognition camera will have other applications and can also be used to detect other violations in public buses, including littering, according to the RTA.