SAFR® is the world’s foremost facial recognition platform for live video. With highest effective accuracy tuned for rapid performance, the fast-growth startup’s computer vision solution — from the established U.S.-based company RealNetworks — is an industry leader in the challenging real-world conditions where faces are in motion, at varying angles, under poor lighting, or partially obscured. Having been designed to operate with a database of over 5M to instantly detect and match millions of faces in a fraction of a second, SAFR’s rapid-performance, highly scalable model for industrial-grade facial recognition answers the call outlined in recent published government RFQs for facial recognition systems to process millions of transactions against a comparably high-volume database.
SAFR offers flexible SDKs that can be easily integrated with edge and IoT devices as part of an industrial-grade solution for world-class face and person detection, recognition, and characterization — including video doorbells, drones, autonomous vehicles, ground robots, and more. SAFR’s high-performance computer vision turns virtually any IP camera into an intelligent lens, unlocking a wide array of user benefits with complete data security and a predictable pricing model. SAFR SDKs can be deployed at the edge or completely offline with no connection to the cloud, offering the same high accuracy, speed, and low bias included in all SAFR deployments.
The SAFR Platform
The foundation for all SAFR solutions is the SAFR platform. Optimized for excellence in live video, the premier facial recognition platform recognizes faces moving through camera feeds even when images are in motion, obscured, tilted, or dimly lit. The high-performance platform delivers 99.87% accuracy in under 100 milliseconds. SAFR is architected to economically scale to thousands of cameras in a distributed architecture, hosted on premises or in the cloud. The platform also provides actionable data for live analytics of traffic volumes, demographic composition, dwell times, and data exports for further reporting. All data passed through the platform is protected by AES-256 encryption, in transit and at rest.
The SAFR platform has 24/7 on-call global coverage, and offers Tier 2 and Tier 3 support. RealNetworks leveraged its 25-year legacy to create SAFR, and has a proven track record of carrier-grade and government-grade services: Its A2P messaging service, KONTXT, interconnects carriers and processes billions of text messages throughout the globe with not a single outage for over 10 years.
The administration console for the SAFR platform, SAFR Desktop makes it possible to connect to a camera, process video, detect faces, and submit images to the SAFR Server or the cloud for recognition.
SAFR Actions provides tangible ways to create and manage responses to event triggers in real time. No matter the use case — multi-factor authentication, building system integration, unobtrusive monitoring, threat detection — SAFR Actions can be written in any scripting language and deployed for a wide range of if-this, then-that scenarios: unlocking a door, turning on a light, sending an alert, sounding an alarm, dispatching police, or any response to fit the use case.
The trusted engine of all SAFR solutions, the SAFR Server includes the facial recognition server, identity database, recognition event server, event archive, report server, and remote video feed administration servers.
SAFR Mobile apps support face detection and facial recognition, as well as provide UI feedback and allow for special use cases, such as setting up a registration kiosk, unlocking doors or adding users — all from a mobile device. SAFR Mobile for Android is available on Samsung Galaxy and Google Pixel devices, and SAFR Mobile for iOS is available on iPads and iPhones.
Access the full capabilities of the SAFR platform through RESTful APIs on all major platforms and operating systems. The SAFR SDK offers face and person detection, recognition, and tracking for live video feeds, age and gender characterization, sentiment analysis, identity management, and event generation and reporting. The SAFR SDK can be deployed on-prem or in the cloud.
Optimized for embedded systems and camera-equipped devices with resource constraints as well as those that lack connectivity, the SAFR Embedded SDK can run entirely offline with no connection to the cloud. Providing an unprecedented blend of accuracy, speed, and efficient size, the SAFR Embedded SDK offers image-based face detection and recognition, identity management, and CLI access to image and database parameters to Android and Windows edge devices.
SAFR Beam is a utility that allows Android users to configure other devices for secure SAFR operation.