Deployable globally or as a managed service, Cydonia is a game-changer for travel and border programmes, digital identity, and verification initiatives across a number of industries such as Government, banking, entertainment events, and retail. Using advanced AI-based facial recognition technology supporting both streaming video and still imagery, Cydonia offers a hugely scalable and cost-effective solution, designed and optimised to run on cloud.

At the foundations of Cydonia lies 'Jezero,' our Secure Cloud Platform, built to be agnostic of any biometric algorithm for maximum flexibility and choice. The platform provides a pre-configured Kubernetes cloud platform, hardened to meet both national security and global compliance standards. Architected leveraging AWS LZA, it not only delivers fundamental value-added services but also includes a cloud operating model, technical standards, and pipeline tooling.

As a biometric vendor agnostic integration company, IdentityE2E carried out benchmark reviews and live trials with AWS, to identify the best face matching system to integrate with our Secure Cloud Platform with which to provide the Cydonia solution. Based on our independent biometric matching performance evaluation and the proven scalability of Paravision’s Image Search Platform to operate on our Secure Cloud Platform, we selected Paravision Search, as the first integration with our platform. It incorporates Paravision’s latest facial recognition algorithms, which have been independently evaluated by the US National Institute of Standards and Technology’s (NIST’s) Face Recognition Technology Evaluation (FRTE) across more than 5 years of consistent submissions, repeatedly validated as being among the most accurate available. The seamless integration of Paravision Search means that Cydonia delivers a highly performant, scalable, enterprise-grade face matching solution.

Features

Cydonia offers unparalleled capabilities, enhances operational efficiency, security, and delivers:

Choice of deployment: Offered as a fully hosted SaaS service or deployable service to customers AWS accounts.

Massive scalability: Cydonia can handle virtually unlimited concurrency and support for galleries of up to hundreds of millions of records while maintaining its NIST-tested accuracy.

Extreme speed: Enrol and match identities at staggering speeds to improve process efficiency and decrease user wait times. Optimised at the assembly language layer, Cydonia delivers over a million matches per second per core.

Native elasticity: Cydonia is fully elastic, allowing programmatic adjustments in compute resources to help meet real-time demands and drive cost savings during low traffic times.

Flexibility: Architecture to help address multiple use case needs from large-scale Identity Verification (IDV) to day-of-travel passenger facilitation, with deployment flexibility and modular features.

Sophisticated gallery management: Works as a distributed database for biometric and vector data, allowing speedy addition and search of data, and enables definition of sub galleries and focused search with adaptive filters.

Lightning-Fast Market Penetration: Cydonia accelerates integration, helping organisations swiftly implement the system with minimal effort.

Cost Optimisation: Optimise your SLAs around peak periods based on workloads and auto scale up and down to only use the compute you need, saving significant cost and improving sustainability.

Presentation Attack Detection Ensuring authentic identity in the physical world iBeta Level 1/Level 2-certified, fully passive liveness detection works on single frame face images, and is optimized for remote, selfie-based enrolment or authentication.

Deepfake Detection Authentic identity in the digital domain Highly accurate single-frame deepfake detection - the ideal complement to presentation attack detection. Optimized for a range of Expression Swap and Identity Swap deepfakes.

Cydonia delivers these benefits with exceptional performance:

• Gallery size scales to hundreds of millions of records • Enrols 50 million identities in less than 3.5 hours • 1.4M matches per second per core (Intel(R) Xeon(R) @ 3.10GHz) • 120 concurrent requests per second (16 CPUs, 100,000 templates) • Optimized for Paravision Face recognition, repeatedly ranked #1 globally in NIST Facial Recognition Vendor Tests (FRVT)

Note We believe in the responsible use of AI. Notwithstanding the current and potential benefits that this solution can bring to society at large, there are also concerns about potential misuses of it, either due to their technical limitations or ethical considerations. Before this solution can be licensed and provisioned, your use of the technology will need to be clearly defined and go through a stringent approval process to ensure it complies with IdentityE2E's use case standards and Paravision’s AI Principles, available on request.