By the SightShield team · · 3 min read
Detection is not recognition
Many cameras advertise “face detection”, which only finds faces in the image and saves a close-up snapshot. Face recognition goes further: it converts each face into a mathematical template and compares it with a database of enrolled people to say who it is. The difference matters both technically — recognition needs better cameras and placement — and legally, because recognition processes biometric identity.
Where businesses use it
| Use | How it works | Consent position |
|---|---|---|
| Staff access and attendance | Enrolled employees are recognised at a terminal or door | Employees informed and enrolled; clear policy |
| Visitor management | Visitors register and are recognised on return visits | Visitor agrees at registration |
| VIP or member recognition | Hotels and clubs greet known guests | Guests opt in |
| Watchlist alerts | Known offenders flagged to security | Sensitive; strict purpose, accuracy and legal review needed |
| Face capture for evidence | Best-quality face snapshots stored with recordings | Ordinary CCTV signage and policy |
Accuracy in the real world
- Recognition works best at controlled points — doors, turnstiles, reception — where people face the camera at close range.
- Wide-angle cameras covering a hall rarely recognise people reliably.
- Masks, caps, helmets, backlight and low light cut accuracy sharply.
- Large watchlists increase false matches, so a human must verify every alert before acting.
- Lighting and camera height should be planned specifically for faces.
Privacy law in India
The Supreme Court recognised privacy as a fundamental right in the Puttaswamy judgment of 2017. The Digital Personal Data Protection Act, 2023, with Rules notified in November 2025 and most duties expected from around May 2027, governs digital personal data, including face images and templates. Businesses need a lawful basis — usually consent or a recognised legitimate use such as employment purposes — clear notice of purpose, security safeguards, limited retention, and a way for people to raise grievances. Penalties for failing to protect data can reach Rs 250 crore.
The DPDP Act and CCTV →Good practice for businesses
- Use recognition only where it solves a defined problem, not everywhere.
- Tell people clearly through signage, notices and policies.
- Enrol only those who need it, and remove templates when they leave.
- Store templates on secured devices or servers with restricted access.
- Keep humans in the loop: never act on an automatic match alone.
- Document the purpose and retention period, and review it yearly.
Face recognition at doors and attendance
The most common and least controversial use is face recognition terminals for employee attendance and door access: people know they are enrolled, the purpose is clear, and the terminal works at close range with high accuracy. Touch-free terminals also replaced many fingerprint readers after 2020.
Face vs fingerprint attendance →Planning a face recognition point
- Camera height around head level, angled no more than slightly downward, so faces are seen nearly front-on.
- Even lighting on faces; avoid a bright glass door behind the person.
- A natural pause point — a door, turnstile or reception desk — so people face the camera for a moment.
- Enrolment photos taken in similar light to the real installation.
- A separate fallback (card or PIN) for people who cannot or do not wish to use face recognition.
Our approach
We install face recognition where it is reliable and proportionate — attendance, access control and visitor management — and design camera height, lighting and enrolment for accuracy. For broader surveillance uses, we recommend face capture with human review rather than automatic identification.
Biometric attendance and access →Quick answers
Is face recognition CCTV legal in India?+
It is not banned for private businesses, but face data is personal data under the DPDP Act, so purpose, notice, security and retention rules apply.
How accurate are face recognition cameras?+
Very accurate at close range at doors and terminals; much less reliable from wide-angle cameras in crowded spaces.
Do I need consent to use face recognition for staff attendance?+
Employers should inform staff clearly and follow DPDP requirements; employment purposes are a recognised basis, but a clear policy is essential.
What is the difference between face detection and recognition?+
Detection finds faces in an image; recognition identifies whose face it is.