Skip to content

RCC · Recognition

Face recognition with privacy controls built in

Match faces against your watchlists in real time, search recorded video by photo, and verify identity at doors and visitor desks, with consent records, approvals and an audit trail on every use.

A face is detected, matched against an approved watchlist, and countersigned before it becomes an alertFour stages run left to right: detection at the edge, matching against an approved watchlist at its threshold, a privacy officer’s countersign for sensitive lists, and the alert with its SOP. Below, three outcomes: an alert, a match discarded below threshold, or a countersign still pending.detectedge servermatchapproved watchlistcountersignprivacy officeralert + SOPwith the clipalert raisedbelow threshold: discardedcountersign pending
  1. Faces are found in the live streams by edge servers next to the cameras.
  2. Each face is compared with approved watchlists, at the threshold set for each list.
  3. Sensitive uses need a privacy officer’s countersign before an alert reaches operators.

Capabilities

What face recognition in RCC does

  • Enrolment and galleries

    Enrol one person or thousands, organise galleries and keep one person database.

  • Watchlists with approvals

    Thresholds, severity and an approval step for each watchlist.

  • Live match alerts

    Matches reach the operator console and the incident queue with video attached.

  • Search and compare

    Search recorded video by photo, and compare two faces one to one.

  • Attendance

    Attendance from faces, with consent recorded for each person.

  • Face-verified entry

    Confirm that the person at the door is the cardholder.

  • Visitor screening

    Screen visitors at check-in, with a countersign for sensitive checks.

  • Privacy by design

    Consent register, countersign, retention per gallery, data subject requests and a tamper-evident log of every search and match.

In the product

Watchlists bound to each camera

Bind a camera to the watchlists it checks, with a class of grant, alert or observe, a match threshold and margin, and a schedule. Consent, erasure and audit sit beside the bindings.

RCC Studio face recognition bindings: cameras at the entrance, lobby, canteen and yard of site HQ, each bound to a watchlist such as staff, VIP guests or a blacklist, with a class of grant, alert or observe, a threshold and margin, a schedule and its state.

How it works

From enrolment to a confirmed match

  1. Enrol

    Add people to galleries, with consent recorded where required.

  2. Detect

    Edge servers find faces in the live streams.

  3. Match

    Faces are compared with approved watchlists.

  4. Respond

    A match opens an alert, an incident and its SOP.

Where it is used

Where face recognition is used

  • Campus safeguarding

    Authorised pickups and a fast response when a child cannot be found.

  • Hospitals

    Restricted wards, pharmacies and maternity units.

  • Industrial sites

    Contractor checks at the gate and hazardous-area entry.

  • Public safety

    Investigations across city cameras, under the agency’s control.

On the RCC platform

Stronger next to the rest of RCC

A watchlist match reaches the incident queue with its video attached, a face confirms the cardholder at the door, and every search is recorded in the same audit trail.

Questions

What buyers ask

Does face data leave our premises?

No. Matching runs on your own servers, on premises or air-gapped, and face data stays in your RCC deployment.

How is misuse prevented?

Watchlists need approval, sensitive searches need a privacy officer’s countersign, and every enrolment, search and match is kept in a tamper-evident log.

Can it run without RCC video management?

Yes. It runs on its own, and adds more value next to RCC video, access control and incident management.

Review the privacy controls with us

Tell us where you want to use face recognition and what your policy requires. We will show the consent, approval and retention controls before any match is made.