Why it exists
Entrances become checkpoints that never lose concentration
Enrolled databases of persons of interest, banned individuals, VIP customers and staff are matched against live video, and the right people are notified in seconds with a photo pair.
Chasing Sun deploys facial recognition with a posture that puts compliance first: we advise on signage, retention settings and policy alignment with applicable data protection law, including Kenya's Data Protection Act, as part of every engagement. Legal review remains the client's responsibility; this is deployment guidance, not legal advice.
Demo footage: official TRASSIR product page
Capabilities
What the module does, as TRASSIR publishes it
- Face detection from any angle, including profiles
- Identification against a photo database with degree of match scoring and quality assessment
- Recognition of appearance attributes (hair color, glasses, headwear) and search by them
- Spoofing defense: detects photographs being shown to the camera
- Face search across the video archive
- Integration into access control systems with selection of the best quality frame
Responsible deployment
Powerful tools deserve mature governance
Proportionality
Recognition where it answers a real risk, not blanket coverage for its own sake.
Transparency
Clear signage and communicated policy wherever recognition operates.
Retention limits
Face data kept only as long as the purpose requires, configured at deployment.
Access control
The face database itself is permissioned, audited and restricted to named roles.
For cooperative identity built on enrollment, such as doors, time and attendance, see our biometric solutions
Inside the module
From a face in the crowd to a scored match
The module detects faces from any angle, selects the best frame and compares it against your enrolled databases. A match returns a degree of match score and a quality assessment, and the operator sees the enrolled image beside the live capture. Every recognition is an archived event in TRASSIR VMS, and Face Search extends the same matching across recorded video.
Appearance attributes such as hair color, glasses and headwear are recognized too, and the archive can be searched by them when no enrolled photo exists.
What a deployment needs
- Camera placement: entrance cameras mounted near face height, aimed where people naturally look forward as they pass.
- Lighting: even light on the face; strong backlight at a glass entrance is the classic enemy, and we design around it.
- Compute and databases: neural matching needs proper server compute, and the face database sits inside your VMS, permissioned and audited.
| Use it for | The database you enroll | What happens on a match |
|---|---|---|
| Security watchlists | Persons of interest and banned individuals | An instant alarm to security, with the photo pair as evidence |
| VIP service | Enrolled VIP customers | A discreet notification to the floor team the moment the guest arrives |
| Access scenarios | Enrolled staff | The system drives access control, with spoofing defense against photos shown to the camera |
| Investigations | Any enrolled or captured face | Face Search finds appearances across the video archive |
From pilot to production
Prove it at one entrance before it guards them all
01 · Demo
See live matching on demo cameras, or run a proof of concept with a small enrolled test group.
02 · Scoped pilot
One or two entrances on your own cameras, an enrolled test database and agreed success criteria.
03 · Tuning
Camera angles, lighting and match thresholds are refined on site until matches are dependable.
04 · Rollout
Extend to further entrances with governance configured before go live: signage, retention and database access.
Questions we hear
Straight answers before you commit
Does the person have to look straight at the camera?
The module detects faces from any angle, including profiles, and selects the best frame for matching. Placement still matters: we position entrance cameras where a usable face view comes naturally.
Can someone fool it with a photo?
The module includes spoofing defense that detects photographs being shown to the camera, which is what makes it usable in access scenarios.
What image quality does it need?
Recognition needs a reasonably sharp, well lit face in the frame. That is a placement and lighting question more than a camera model question, and it is what a pilot verifies.
Is this compliant with data protection law?
Face recognition is deployable when it is governed properly. We advise on signage, retention settings and alignment with applicable law, including Kenya's Data Protection Act; your own legal review remains part of the project.
Where does the face data live?
In your TRASSIR system, under your control: the database is permissioned, audited and restricted to named roles, with retention configured at deployment.
Can it search footage we already recorded?
Yes. Face Search works across the video archive, so an enrolled photo can be traced back through past recordings, not just live streams.
Discuss a compliant face recognition pilot
Tell us the entrances and the risk you are managing. We'll scope cameras, databases and the governance that keeps it defensible.