Stranger Detection
February 7, 2024
Introducing AI Video Analytics for Stranger Detection: Our system uses smart technology to watch over your space. It can quickly spot unknown people and
Solutions · AI Solutions
One deep learning model, trained to find and recognise faces in a camera feed, quietly does three jobs across Collie: it confirms who someone is, lets them through a gate without touching anything, and logs the moment they arrived — all from the same enrolment.
Why it matters
Checking an ID card by eye, or a card by hand, is slow, easy to fool, and forces the same identity to be enrolled separately in every system that needs it.
A guard visually checking IDs against a photo works for a handful of people — not hundreds moving through a gate at peak hours.
Cards and fingerprint scanners mean everyone touches the same surface. A face requires no contact at all.
Without a shared identity engine, the same person ends up re-enrolled separately for access, attendance, and verification.
What’s inside
Two distinct steps, both handled by the same engine.
Locates any face in frame — regardless of angle, distance, or partial obstruction — as the first step before any identity is checked.
Compares a detected face against the enrolled database in real time, returning a match with a confidence score.
Enrol a face once, and it’s recognised everywhere Collie checks identity — no separate registration per system.
Trained models handle lighting changes, angle, and everyday variation far better than simpler template-matching approaches.
How it works
Video analytics can run on your existing CCTV, NVR, or VMS — Collie’s platform adapts to the infrastructure you already have.
A camera at the gate, kiosk, or device locates a face in the frame the moment someone steps into view.
The detected face is compared against the enrolled database, returning a match with a confidence score in real time.
A confirmed match unlocks whatever the situation calls for — a turnstile, a clock-in, or a verified identity check.
One engine, several uses
Because detection and recognition live in one shared engine, the same enrolled identity works everywhere Collie checks who someone is.
Confirms someone is who they claim to be — at a kiosk, a checkpoint, or an ID lookup.
Unlocks a turnstile or gate the instant a match is confirmed, no card or badge required.
Logs a clock-in the moment a face is matched, feeding straight into attendance reporting.
Where it’s deployed
Cardless door access and daily attendance from the same enrolled face.
Worker and contractor verification at gates without a shared touchpoint.
Staff verification at sensitive areas, paired with visitor sign-in.
Fast, contactless check-in for large groups arriving in a short window.
Face Detection is the underlying AI capability that identifies and matches faces in a video feed; Access Control Management is one application of it, used specifically to open gates and turnstiles.
Yes, the system is designed to recognise a registered face even when partially covered by a mask, reducing failed matches at entry points.
Yes – an unmatched face is flagged as unknown, which is the basis for the separate Stranger Detection feature that raises an alert for unrecognised individuals.
Facial data feeds into the same access-controlled Collie system covered by Magicsoft’s ISO 27001-certified information security practices.
It’s used across offices, hospitals, industrial sites, and construction sites – anywhere contactless, fast identity verification replaces a manual ID check.
Introducing AI Video Analytics for Stranger Detection: Our system uses smart technology to watch over your space. It can quickly spot unknown people and
Tell us where you need identity checked — a gate, a kiosk, a clock-in — and we’ll show you how one enrolment covers it all.