UWC tests facial recognition system on campus CCTV, considers full rollout

The University of the Western Cape tested a facial recognition system integrated into its existing CCTV infrastructure
The Swiss-South African firm SAFEZA deployed the technology across three main camera network streams covering thousands of students daily
UWC is now considering a full rollout that could replace student cards and flag security threats automatically at campus gates

UWC tests facial recognition on its CCTV and students are already talking — University of the Western Cape tests tech. Image: Klaus Vedfelt
Source: Getty Images

The University of the Western Cape (UWC) has tested a facial recognition security system across its Bellville campus, with more than 23,000 students on its rolls. The pilot was conducted in partnership with Swiss-South African technology firm SAFEZA.

The trial ran for five consecutive calendar days, covering three working days and one weekend. What made the rollout notable was how quickly it came together — SAFEZA integrated the system into UWC's existing CCTV infrastructure in roughly two hours, using a secure Tailscale mesh VPN connection built on the WireGuard protocol.

How the technology works

Rather than matching faces to photographs, the system converts each person's facial features into a mathematical representation called a facial vector. These vectors are what the recognition engine uses to verify identities, and they are treated as sensitive personal data. AI inference for the system was handled by Johannesburg-based company Gigabrain, meaning all processing stayed within South Africa rather than being routed to an overseas server.

Three of the university's main camera network streams fed into the system throughout the pilot, processing the faces of students entering the campus each day. SAFEZA reported no noticeable delays in identification, despite the AI engine being based hundreds of kilometres away in Gauteng.

What a full rollout could look like

Among the potential applications UWC discussed was a watchlist function that would automatically flag individuals barred from campus, such as former employees or persons identified as security risks. Security personnel would be alerted in real time and could then remove those individuals from the grounds.

The system could also eliminate the need for physical student cards at access gates. SAFEZA said in its case study:

"Instead of requiring every authorised individual to stop, locate a card and physically scan it, facial recognition could support identity verification and automatically record entrance and exit events."

Bently Ebrahim Jeftha, UWC's director of risk and compliance, was positive about the experience. He said:

"The facial recognition functionality performed well during the POC, with accurate identification results and a very low incidence of false readings."

He described the setup process as "effortless."

Once the pilot concluded, all data collected during the trial was deleted and the system was decommissioned. SAFEZA confirmed that participants in the pilot were informed about the technology in advance.

Discussions about a permanent deployment are still ongoing. A full rollout would require all students and staff to enrol their facial vectors into the system. Extended features under consideration include live weapon detection, number plate recognition and driver-vehicle association.

Read the full MyBroadband report here.

UWC tests facial recognition system on campus CCTV, considers full rollout

How the technology works

What a full rollout could look like

More Briefly News stories on universities