Face Recognition
A face recognition system that identifies and verifies people from live video in milliseconds for access control, attendance and security.
Request a demo
The Alfabeta face recognition system detects, aligns and matches faces from live camera feeds or still images against your enrolled identity database. Built on deep neural networks tuned for real-world conditions, it handles variations in pose, lighting, expression and partial occlusion such as glasses or masks, and can detect and track multiple faces in the same frame simultaneously. Every recognition event is logged and time-stamped, giving you a reliable audit trail for facial recognition attendance, investigations and reporting.
The engine supports both 1:1 verification, which confirms a claimed identity, and 1:N identification, which searches a face against your entire enrolled population. It runs on edge devices or centralised servers so you can balance latency, cost and data residency, and it ingests standard camera streams over RTSP, USB or other common protocols. Because the technology is delivered as an API, it drops into access control gates, turnstiles, kiosks and existing VMS platforms with minimal custom development.
Accuracy and speed are central to the design. The models are tuned to minimise false positives while keeping matching fast enough for frictionless, touchless entry, so people are recognised as they walk toward a door rather than stopping to badge in. Alfabeta already processes more than 400,000 active users per month across deployments, and the same core powers related products such as the Alfa Gate Management System (AGMS), the Alfa Visitor Management System (AVMS), attendance recording and payment authentication.
For deployments that demand strong assurance, the face recognition system pairs with Alfabeta Liveness Detection to block photo, video and mask spoofing before a match is trusted, and it complies with data privacy regulations through configurable data retention policies. Pricing follows an adjustable tiered model based on monthly API call volume, so cost scales with actual usage rather than fixed licences, and the system integrates cleanly with your existing infrastructure whether you run on-prem or in the cloud.
Real-Time Matching
Detects and identifies faces from live video in milliseconds, enabling frictionless touchless access and continuous monitoring. Every event is logged and time-stamped for audit and reporting.
1:1 and 1:N Modes
Verify a claimed identity one-to-one or search a face against your entire enrolled database from a single engine. The same models cover both access control and watchlist scenarios.
Robust In The Wild
Maintains accuracy across changes in angle, lighting, expression and partial occlusion like glasses or face masks. Deep-learning models are tuned to minimise false positives in real-world conditions.
Edge Or Cloud Deployment
Runs on embedded edge devices or centralised servers so you can balance latency, cost and data residency. It ingests standard camera streams over RTSP, USB and other common protocols.
Easy Integration
Connect to access controllers, attendance systems and VMS platforms through clean, well-documented APIs. The same technology powers AGMS, AVMS and payment authentication for seamless roll-out.
Liveness-Ready
Pairs with Alfabeta Liveness Detection to block photo, video and mask spoofing before a match is trusted. Configurable data retention policies help you stay aligned with privacy regulations.
Frequently asked questions
How accurate is the face recognition system?
The engine uses deep neural networks tuned for real-world conditions and is designed to minimise false positives while sustaining high match rates across pose, lighting, expression and partial occlusion such as glasses or masks. Accuracy in your environment depends on camera placement, resolution and enrolment image quality, and we tune thresholds during deployment to fit your security and convenience requirements.
Can it integrate with our existing access control and attendance systems?
Yes. The technology is delivered as a documented API and event stream, so it connects to door controllers, turnstiles, kiosks, VMS platforms and attendance software without replacing your existing hardware. It also underpins the Alfa Gate Management System (AGMS) and Alfa Visitor Management System (AVMS) if you prefer a ready-made stack.
Can we deploy on-premise instead of the cloud?
Yes. The system runs on edge devices or centralised on-premise servers as well as in the cloud, letting you keep facial data inside your own network for latency, bandwidth or data-residency reasons. It ingests standard camera feeds over RTSP, USB and other common streaming protocols.
How is it priced?
Pricing follows an adjustable tiered model based on your monthly API call volume, so cost scales with actual usage rather than a fixed per-seat licence. Our team sizes a tier around your expected traffic and can adjust it as your deployment grows.
How do you handle spoofing and data privacy?
The face recognition system pairs with Alfabeta Liveness Detection to block photo, video replay and mask attacks before a match is accepted. For privacy, recognition data is logged and time-stamped, and configurable data retention policies let you align storage with the regulations that apply to your organisation.